Chemmanur Et Al 2019

https://doi.org/10.
1017/S0022109018001497
Downloaded from https://www.cambridge.org/core. Boston College, on 01 Jul 2019 at 00:49:45, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
JOURNAL OF FINANCIAL AND QUANTITATIVE ANALYSIS Vol. 00, No. 00, 2019, pp. 1–40
COPYRIGHT 2018, MICHAEL G. FOSTER SCHOOL OF BUSINESS, UNIVERSITY OF WASHINGTON, SEATTLE, WA 98195
doi:10.1017/S0022109018001497
Top Management Human Capital, Inventor

Mobility, and Corporate Innovation
Thomas J. Chemmanur , Lei Kong, Karthik Krishnan,
and Qianqian Yu*
Abstract
Using panel data on top management characteristics and a management quality factor con-
structed using common factor analysis on individual management quality measures, we
analyze the relation between top firm management quality and corporate innovation input
and output. We show that top management quality is an important determinant of corporate
innovation, with individual aspects of management quality affecting innovation in younger
and older firms differently. Further, firms with higher top management quality engage in
more risky (“explorative”) innovation strategies. Finally, hiring more and higher-quality
inventors is an important channel through which firms with higher management quality
achieve greater innovation output.
I. Introduction
The effectiveness of a firm’s top management team in investing in and man-
aging innovative projects may determine the long-term success of the firm. In-
deed, prior literature suggests that firms’ investments in research and development
(R&D) and their innovative output (measured by patents and citations) may have
a positive impact on the long-term financial health of the firm. Given this, there
is surprisingly little analysis of how the human capital or “quality” of a firm’s top
*Chemmanur (corresponding author), chemmanu@bc.edu, Boston College Carroll School of

Management; Kong, lkong@cba.ua.edu, University of Alabama Culverhouse College of Business;
Krishnan, k.krishnan@northeastern.edu, Northeastern University D’Amore–McKim School of Busi-
ness; and Yu, qiy617@lehigh.edu, Lehigh University College of Business and Economics. We thank
Paul Malatesta (the editor) and an anonymous referee for several valuable comments that helped to
greatly improve this paper. We are grateful for comments and suggestions from Kee H. Chung, Marcia
Millon Cornett, Michael Ewens, Chinmoy Ghosh, Joseph Golec, Shantaram Hegde, Sahn-Wook Huh,
Brad Jordan, Andrew Karolyi, Mark Liu, Gustavo Manso, Alan Marcus, Kristina Minnick, Kartik
Raman, Bob Taggart, Hassan Tehranian, Brian Wolfe, and Zhaoxia Xu; seminar participants at Boston
College, Bentley University, the University of Kentucky, SUNY Buffalo, the University of Calgary,
the University of Connecticut, and Hong Kong Polytechnic University; and conference participants at
the 2015 PBC-RFS conference on “Entrepreneurial Finance and Innovation around the World,” the
2015 Financial Management Association Annual Meeting, and the 2016 American Finance Associa-
tion Meetings in San Francisco.
1
https://doi.org/10.1017/S0022109018001497
2 Journal of Financial and Quantitative Analysis
management team may affect its innovation input and output. We aim to fill this
gap in the literature.
One strand in the theoretical literature suggests that higher quality manage-
ment teams may invest in a larger number of value-enhancing long-term projects
(e.g., Chemmanur and Jiao (2012)). Given that innovative projects are among such
long-run value-enhancing projects (e.g., Hirshleifer, Hsu, and Li (2013), Griliches
(1990)), we expect that higher-quality top management teams will invest in more
innovative projects and will have a greater extent of innovative output, on average.
They may accomplish this by having better foresight into the potential value of
innovative investment opportunities and by more effectively managing innovative
resources such as physical assets (e.g., research equipment) and human capital
(e.g., scientists and inventors). Further, they may provide an environment that
fosters greater failure tolerance in the sense of Manso (2011). Given this, firms
with higher-quality management teams may attract inventors with greater skills to
work for them.
These arguments lead to several testable predictions. First, firms with higher-
quality top management teams may invest more in R&D. Second, firms with
higher quality top management teams may have a greater extent of innovation
productivity (measured by the number of patents) and higher-quality innovation
(measured by total citations and citations per patent). Third, better management
of innovative assets by higher-quality management teams may be reflected in a
greater extent of innovative efficiency (e.g., patents per R&D dollar) of the firms
under their management. Fourth, the effect of management team quality on inno-
vative output may be stronger for firms facing financial constraints and for firms
in competitive industries. Since such firms are at a disadvantage relative to other
firms, the “leg-up” provided by higher-quality top management teams may en-
hance their future prospects.
It is also likely that different aspects of top management team quality may
have significantly different effects on the innovation output of firms at different
innovation stages: For example, younger firms may have a larger fraction of their
innovation projects at the early innovation stage, whereas older firms may have
a larger fraction of their innovation projects at the commercialization stage, so
that the individual aspects of the top management team that are most important to
efficiently manage innovation in these two types of firms may be different. The
top management quality of a firm may also affect its innovation strategy. On the
one hand, it is possible that, given their greater human capital, firms with higher
top management quality will pursue more risky innovation strategies involving
new technologies and are more likely to push their knowledge boundaries out-
ward. However, it is also equally possible that such firms may engage in less risky
(more conservative) innovation strategies. Finally, an important channel through
which firms with higher-quality top management teams achieve greater innova-
tion output may be by hiring a greater number of inventors (for a given level of
R&D expenditures) and by hiring higher-quality inventors (as measured by their
prior track record of citations per patent).
An empirical analysis of the relationship between top management qual-
ity and corporate innovation faces two challenges. First, measuring the human
capital of a firm’s top management team (which we refer to as top management
https://doi.org/10.1017/S0022109018001497
Chemmanur, Kong, Krishnan, and Yu 3
quality) involves subjective notions of what constitutes a higher-quality manage-

ment team. Second, potential endogeneity may confound empirical findings on
the relation between management quality and innovation. In particular, there may
be an endogenous matching between higher-quality top management teams and
higher-quality firms.
We overcome the first challenge above by constructing a management team
quality factor from seven measures capturing various individual aspects of top
management quality following the existing literature (see, e.g., Chemmanur and
Paeglis (2005), Chemmanur, Paeglis, and Simonyan (2011)): management team
size, the fraction of managers with MBAs, the fraction of managers with Ph.D.s,
the fraction of members with prior work experience in the top management team,
the average number of prior board positions that each manager has served on, and
the average employment- and education-based connections of each manager on
the management team.1 These measures are adjusted for firm size. We create our
index of top management quality using common factor analysis to capture the co-
movement of the aforementioned seven individual measures of top management
quality, thereby extracting a single “management quality factor.”2 To test the va-
lidity of this “management quality factor” as a measure of the quality of a firm’s
management team, we analyze the relation between our management quality fac-
tor and the average compensation of senior managers included in our management
quality measure, and we observe a strong and significantly positive correlation
between these two variables, thus establishing the validity of our management
quality factor.
We overcome the second challenge related to endogeneity by using an in-
strumental variable (IV) analysis. We instrument for top management quality (as
measured by our management quality factor) using a plausibly exogenous shock
to the supply of top executives available for hire by a firm, namely, the number of
acquisitions in the industry and state of the sample firm 5 years prior, weighted
by an index measuring the enforceability of noncompete clauses in that state
1
Starting with the pioneering work of Becker (1964) and Ben-Porath (1967), the labor economics
literature has focused on the human capital of workers. The Becker view is that human capital increases
a worker’s productivity in all tasks, although possibly differentially in specific tasks, organizations, and
situations. In the Becker view, although the role of human capital in the production process may be
quite complex, it is representable by a unidimensional measure, such as a worker’s stock of knowledge
or skills, which becomes directly part of the production function. When analyzing the human capital
of the members of a firm’s top management team, our view is that managerial human capital is mul-
tidimensional, consisting of many different aspects, which we capture using the individual measures
we mention here and collapse into one factor using common factor analysis. Thus, our view of human
capital is closer to the view of the social psychologist Howard Gardner (see, e.g., Gardner (1983) and
Acemoglu and Autor (2011) for a review). An advantage of such a multidimensional approach is that
it allows us to capture differences in not only the quantity but also the quality of the human capital of
top management teams across firms.
2
Chemmanur and Paeglis (2005) and Chemmanur et al. (2011) use a broadly similar approach to
study the effects of the management quality of private firms on the initial public offering (IPO) char-
acteristics of these firms and on the anti-takeover provisions in their corporate charters, respectively.
Chemmanur, Paeglis, and Simonyan (2009) also use a similar approach to analyze the relationship
between the management quality and financial policies of firms making seasoned equity offerings
(SEOs). Unlike the current paper, the previously mentioned papers make use of cross-sectional data
hand-collected from IPO and SEO prospectuses.
https://doi.org/10.1017/S0022109018001497
(aggregated to the national level). We discuss the rationale for this instrument
in more detail in Section VI.
We analyze the relationship between top management quality and corporate
innovation using a panel data set of 4,389 firms covering the period 1999–2009.
We obtain biographical data on the top managers of firms from the BoardEx
database, patent and citation information from the patent data set created by
Kogan, Papanikolaou, Seru, and Stoffman (2017) based on data from the United
States Patent and Trademark Office (USPTO), and inventor information associ-
ated with each patent from the U.S. Patent Inventor Database (1975–2010); see
Li, Lai, D’Amour, Doolin, Sun, Torvik, Yu, and Fleming (2014) for a detailed
description of the latter database.
Our empirical results may be summarized as follows: First, we find that
higher-quality management teams invest more in R&D, showing that they de-
vote a larger amount of resources to innovative activities (i.e., have larger input
into corporate innovation). Second, firms with higher-quality management teams
have a greater extent of innovation output (measured by the number of patents)
and higher-quality innovation output (measured by total citations and citations
per patent). These effects are also economically significant. For instance, a 1 in-
terquartile range increase in top management quality increases firm patents by
12.8%. Third, we find that firms with higher-quality management teams produce
more patents and citations per R&D dollar, that is, have greater innovative ef-
ficiency (see, e.g., Hirshleifer et al. (2013)). The aforementioned results on the
relation between management quality and corporate innovation are confirmed by
our IV analysis making use of the instrument mentioned previously. Finally, the
relation between top management team quality and innovation is stronger for firms
in financially constrained industries and those in more competitive industries.
We also analyze how the seven individual measures discussed previously
that capture various individual aspects of top management quality affect corpo-
rate innovation. We find that most of these individual measures have a positive
and significant effect on the quantity and quality of corporate innovation. We then
analyze how these individual measures for top management team quality affect the
innovation output in younger firms (likely to have a larger fraction of innovative
projects at the “early innovation” stage) versus older firms (likely to have a larger
fraction of innovative projects at the commercialization stage). We find that man-
agement team size, the fraction of managers with MBA degrees, and education-
based connections positively and significantly affect corporate innovation only for
older firms. Prior managerial work experience positively and significantly affects
corporate innovation only for younger firms. Two individual measures, namely,
the fraction of top managers with Ph.D. degrees and employment-based connec-
tions, positively and significantly affect corporate innovation in both younger and
older firms; however, the magnitude of the effect of the Ph.D. measure is greater
for corporate innovation in younger firms.
We also examine the relation between the nature of the innovation strategies
undertaken by firms and their top management quality. In particular, we analyze
whether firms with higher-quality top management teams engage more in risky
“explorative” innovation strategies (where they venture into the development of
newer technologies or pursue innovations in areas that are less familiar to the firm)
https://doi.org/10.1017/S0022109018001497
or in more conservative “exploitative” innovative strategies (where they may pur-

sue innovations using more conventional technologies or in areas that are more
familiar to the firm). We find support for the notion that firms with higher top
management quality engage in riskier innovation strategies, innovations involv-
ing new technologies, and those pushing forward the knowledge boundaries of
the firm. Consistent with this, we find that the relation between top management
quality and the fraction of firm patents using new knowledge is positive and sig-
nificant (and is stronger than the relation between top management quality and the
fraction of firm patents using existing knowledge). Further, top management qual-
ity is positively and significantly related to innovation diversity (see, e.g., Brav,
Jiang, Ma, and Tian (2018)) and to the number of firm patents that belong to the
most highly cited (top 10%) group of patents (see, e.g., Balsmeier, Fleming, and
Manso (2017)).
Finally, we investigate an important channel through which higher-quality
management teams may foster greater innovation in their firms. We argue that
higher-quality management teams may provide more resources to R&D, manage
R&D resources better, and provide a more failure-tolerant climate for inventors
to succeed in. This, in turn, may make firms with higher top management quality
more attractive to higher-quality inventors. Thus, one way that higher-quality top
management teams may enhance innovation is by hiring more and higher-quality
inventors to work for the firm. We find evidence consistent with this hypothesis:
Firms with higher-quality management teams experience greater net inflows of
inventors (controlling for R&D expenditures), particularly of higher-quality in-
ventors (defined as those inventors with a record of past citations per patent above
that of the median inventor in our sample). We also find that the average citations
per patent of incoming inventors entering firms with higher-quality management
teams are higher than the average citations per patent of outgoing inventors from
such firms.
Our paper contributes to several strands in the literature. First, we
contribute to the recent literature that has analyzed the determinants of
innovation in firms theoretically as well as empirically (e.g., Manso (2011),
Marx, Strumsky, and Fleming (2009)) and their impact on firm performance (e.g.,
Hirshleifer et al. (2013), Gu (2005), Eberhart, Maxwell, and Siddique (2004),
Lanjouw and Schankerman (2004), Lerner (1994), and Griliches (1990)). The ex-
isting literature has focused, however, on firm characteristics other than top man-
agement team quality and its effects on innovation. Some of these characteristics
are managerial compensation (e.g., Lerner and Wulf (2007), Ederer and Manso
(2013)), public versus private status (see e.g., Bernstein (2015) for an empirical
analysis and Ferreira, Manso, and Silva (2014) and Spiegel and Tookes (2016)
for theoretical analyses), private equity or venture capital involvement (e.g.,
Lerner, Sorensen, and Strömberg (2011), Tian and Wang (2014), and Chemmanur,
Loutskina, and Tian (2014)), anti-takeover provisions (e.g., Atanassov (2013),
Chemmanur and Tian (2018), and Sapra, Subramanian, and Subramanian (2014)),
institutional ownership (e.g., Aghion, Van Reenen, and Zingales (2013)), CEO
overconfidence (Hirshleifer, Low, and Teoh (2012)), and conglomerate structure
(e.g., Seru (2014)). In a contemporaneous paper, Custodio, Ferreira, and Matos
(2018) analyze how the general versus firm-specific human capital of CEOs
https://doi.org/10.1017/S0022109018001497
affects corporate innovation. In contrast to the last paper just mentioned, our paper
does not focus on CEO characteristics: rather, we analyze the relation between top
management team quality and corporate innovation, thus moving the literature in
a new direction.
Second, we contribute to the literature linking top management quality to
firm performance, investments, and financing. For instance, Bertrand and Schoar
(2003) study the effect of top managers on a firm’s financial and investment
policies.3 They find that manager fixed effects explain some of the heterogene-
ity in the investment, financial, and organizational practices of firms. Ewens
and Rhodes-Kropf (2015) use a similar fixed-effects methodology to investigate
whether individual venture capitalists have repeatable investment skill and the
extent to which their skill is impacted by the venture capital firms for which
they work. They accomplish this by tracking the performance of individual ven-
ture capitalists’ investments over time as they move between firms. Chemmanur,
Kong, Krishnan, and Yu (2017) relate measures of top management quality similar
to ours to stock return, operating performance, and firm valuation. Unlike the last
paper just mentioned, we focus on the relation between top management quality
and corporate innovation output. Further, we add to the above literature by ana-
lyzing the nature of the innovative strategies adopted by firms with higher- versus
lower top management team quality.4 We also provide new evidence suggesting
that higher-quality managers enhance their firms’ innovation output by attracting
higher-quality inventors to join the firm.5
The rest of the paper is organized as follows: Section II discusses the un-
derlying theory and develops testable hypotheses. Section III describes our data
and sample selection procedures, explains the construction of our measures of
top management quality and other measures, and validates our management qual-
ity measure using managerial compensation data. Section IV presents our em-
pirical tests and results. Section V presents our analysis of the inventor mobil-
ity channel. Section VI presents our IV analysis. Section VII concludes. The
Supplementary Material, available at https://www2.bc.edu/thomas-chemmanur/,
develops testable hypotheses regarding the effect of various individual aspects of
management team quality on corporate innovation and presents additional robust-
ness tests not included in the main text due to space limitations.
3
Our paper is also indirectly related to the literature analyzing the determinants of CEO quality and
how it affects firm performance (see, e.g., Adams, Almeida, and Ferreira (2005), Malmendier and Tate
(2005)). See also Kaplan, Klebanov, and Sorensen (2012), who study the individual characteristics of
CEO candidates for companies involved in buyout and venture capital transactions and relate them to
the subsequent performance of their companies.
4
In more distantly related research, Bloom and Van Reenen (2007) use an innovative survey tool
to collect data on management practices from various countries and show that measures of managerial
practice are strongly associated with firm-level productivity, profitability, Tobin’s Q, and survival rates.
5
Two important papers in the economics literature that have implications for our paper are those by
Sah and Stiglitz (1986), (1991), which imply that firms with larger management teams are more likely
to reject bad projects. The organizational behavior literature on the effect of managerial discretion on
firm performance is also indirectly related to our paper: see Finkelstein and Hambrick (1996) for a
review.
https://doi.org/10.1017/S0022109018001497
II. Theory and Hypothesis Development

In this section, we briefly discuss the underlying theory and develop hypothe-
ses for our empirical tests. Our theoretical motivation depends on arguments made
by Chemmanur and Jiao (2012) and Stein (1988). Chemmanur and Jiao (2012)
study a setting in which managers with greater talent or ability are able to gen-
erate greater long-run cash flows by undertaking long-term projects. However,
since their talent is private information, and since short-term projects come to
fruition earlier, myopia or short-termism induced by the stock market (e.g., due
to the possibility of rivals appearing and successfully taking over the firm in the
absence of favorable signals of project success in the short run) impose pres-
sures on them to undertake short-term rather than long-term projects (see also
Stein (1989)). However, Chemmanur and Jiao (2012) show that more capable
managers also have an incentive to undertake long-term (innovative) rather than
short-term projects, since they are able to create greater long-term value by do-
ing so. In such a setting, the equity market values firms undertaking long-term
projects at a discount, since they are not able to observe true managerial abil-
ity; however, firms with higher perceived management team quality suffer only a
smaller valuation discount if they undertake long-term projects. In summary, man-
agers’ choice between long-term and short-term projects is driven by the trade-off
between the pressure induced by a myopic stock market versus the ability (and de-
sire) of more able managers to create greater value in the long run by undertaking
long-term projects.6 Given that innovative projects are long-term projects charac-
terized by short-term failures and experimentation (which increases the gestation
time of these projects), managers with greater perceived ability will undertake
a greater proportion of long-term (innovative) rather than short-term projects in
equilibrium.7
The aforementioned theoretical framework provides us with our first set of
testable implications. First, top management teams with higher (perceived) qual-
ity are likely to devote a greater amount of resources to innovation activities.
Thus, firms with higher quality management teams will be characterized by a
larger input of resources into innovation activities (i.e., R&D expenditures) (Hy-
pothesis 1). Further, we expect such firms to be characterized by greater inno-
vation output and higher quality of innovation output, after controlling for R&D
6
Formally, in Chemmanur and Jiao (2012), the objective function of the manager is a weighted
average of the short-run and long-run stock price. Thus, while talented (higher-ability) managers will
suffer a discount in the firm’s short-run stock valuation if they take a greater proportion of long-
term projects (since their equity will be priced in a pooling equilibrium with firms with less talented
managers), more talented managers have an incentive to undertake a greater proportion of long-term
projects since these projects will allow them to create greater long-run value, yielding a higher long-
run stock price.
7
Note that this theoretical framework does not require the assumption that long-term projects are
always better than short-term projects. In industries where short-term projects are as valuable as long-
term projects (e.g., in less innovative industries), there is no problem of equity undervaluation (since
all managers will prefer to invest in short-term projects in such industries), so that the relation between
management quality and innovation will be weaker in such industries.
https://doi.org/10.1017/S0022109018001497
expenditures (Hypothesis 2).8 Such firms will also be characterized by greater in-
novative efficiency (i.e., greater innovation output and higher-quality innovation
output per dollar of R&D capital investment) (Hypothesis 3).9
So far, we have discussed top management team quality without providing
any detail about how we measure top management quality. In order to develop
testable hypotheses about how various individual aspects of top management qual-
ity are related to corporate innovation, we now briefly discuss the approach that
we take to measure top management quality. We first measure various individual
aspects of the top management team quality of a firm using seven measures pre-
viously used in the literature. These seven individual measures of management
team quality are management team size (TEAM SIZE), the fraction of managers
with MBA degrees (MBA), the fraction of managers with Ph.D. degrees (PHD),
the fraction of managers with prior work experience (WORK EXP), the average
number of prior board positions that each manager has served on (BOARD EXP),
and the average employment- and education-based connections of each manager
in the management team (EMP CONN and EDU CONN). Then, given that true
top management quality itself is fundamentally unobservable, we make use of the
comovement between these seven individual measures (correlated with the un-
derlying unobservable true management quality) to extract a management quality
factor (MQF) using common factor analysis. The details of how these seven in-
dividual aspects of top management team quality are measured are discussed in
Section III.B.1. Due to space limitations, we confine our discussion of how each of
these aspects (measures) of top management quality may individually be related
to corporate innovation and the differential impact of these measures on innova-
tion in younger versus older firms to Section 1 of our Supplementary Material.
As is discussed in detail in the Supplementary Material, we expect each of the
seven individual measures of top management team quality discussed above to be
positively related to the quantity and quality of the innovation output of a firm.
This is the next hypothesis that we test here (Hypothesis 4).
We now delve deeper into the possible differences in the innovation strate-
gies adopted by firms with higher- versus lower top management quality. One
possibility is that firms with higher top management quality pursue more risky
innovation strategies, those involving new technologies, and those that are likely
to push the knowledge boundaries of the firm outwards (Hypothesis 5A). In this
case, we would expect firms with higher top management quality to engage in
more explorative innovation strategies (in the sense of Balsmeier et al. (2017),
8
Since innovation is only one of the many activities undertaken by firms, it is not a priori obvious
that higher top management quality (as captured by our measures) will be associated with greater
innovation output, higher-quality innovation output, or greater innovation efficiency. For example, we
will not find such associations empirically if higher-quality top management teams invest more in
short-term projects (see, e.g., Hirshleifer and Thakor (1992), who argue that managerial concern for
reputation results in conservatism in project choice) or are better at managing the commercialization
of innovations but not at managing the development of these innovations.
9
As in practice, here we are assuming that higher-quality managers are in short supply, so that there
is considerable variation in management quality across firms. Clearly, even when higher-quality man-
agement teams command a larger amount in aggregate compensation, hiring a higher-quality manage-
ment team will be optimal for firms as long as the incremental surplus (shareholder value) generated
by hiring such managers after accounting for their higher compensation is positive.
https://doi.org/10.1017/S0022109018001497
Brav et al. (2018)), so that they venture into the development of newer technolo-
gies or pursue innovation in areas less familiar to the firm. Given that such an
explorative strategy is more risky, under this scenario, we would expect firms
with higher management quality to be associated with a larger number of highly
successful and a larger number of quite unsuccessful innovations (as measured by
citations per patent) compared to firms with lower management quality. In other
words, in this case, firms with higher management quality will have a larger num-
ber of patents in the two tails of the patent quality distribution. Further, in this case,
we would also expect top management quality to have a greater effect on non-self-
citations than on self-citations. Finally, we would expect top management quality
to be positively associated with greater innovation diversity. Alternatively, firms
with higher management quality may engage more in less risky (conservative)
innovation strategies (Hypothesis 5B). In this case, we would expect firms with
higher top management quality to pursue more exploitative innovation strategies
(in the sense of Balsmeier et al. (2017), Brav et al. (2018)), those using more
conventional technologies, and those in areas that are more familiar to the firm.
Further, in this case, we would expect firms with higher top management quality
to be associated with more moderately successful innovations (again measured by
citations per patent) compared to those achieved by firms with lower management
quality.10 Finally, in the latter case, we would also expect top management qual-
ity to have a greater effect on self-citations than on non-self-citations and to be
negatively related to innovation diversity.
We now analyze whether the relationship between top management quality
and corporate innovation productivity is stronger for firms in some industries than
those in others. First, consider firms in financially constrained industries. Given
their financial constraints, such firms will have only a limited amount of resources
to devote to innovation. If the relation between top management quality and inno-
vation is partly driven by more effective resource management on the part of firms
with higher management quality, we would expect the relationship between man-
agement quality and innovation to be stronger for firms in financially constrained
industries (Hypothesis 6).11 Next, consider firms in more competitive versus those
in less competitive industries. Scientists and engineers (inventors) in more com-
petitive industries are likely to have greater outside employment opportunities,
so that talented inventors are likely to be in limited supply in these industries.
Therefore, if firms with higher-quality top management teams are able to attract a
greater proportion of these talented inventors in limited supply, we would expect
the relationship between top management quality and innovation productivity to
be stronger in more competitive industries (Hypothesis 7).
Finally, we analyze an important channel through which firms with
higher top management quality may be able to generate greater innovation
10
It is difficult to predict from a priori theoretical considerations which of the above two scenarios
will be realized in practice. We will therefore leave this question to be resolved empirically.
11
While firms in financially unconstrained industries may be able to partially compensate for not
having higher-quality management teams by devoting more resources to innovative activities (e.g., by
buying higher-quality innovation equipment), firms in financially constrained industries will be less
able to do so, so that the relationship between management quality and innovation will be stronger for
the latter category of firms.
https://doi.org/10.1017/S0022109018001497
productivity (i.e., greater innovation output for a given amount of resources de-
voted to R&D expenditures). Consistent with our conjecture that higher-quality
top management teams may be able to manage their innovative activities more
efficiently, we hypothesize that such firms are able to hire more inventors for a
given amount of R&D expenditures (Hypothesis 8). We further conjecture that
such firms are likely to hire higher-quality inventors, who are more innovative (as
measured by these inventors’ prior track record of citations per patent).12 This is
the final hypothesis that we test here (Hypothesis 9).
III. Data and Sample Selection

A. Sample Selection
Our sample is derived from multiple data sources. Our primary data source
for the biographical information of senior managers is the BoardEx database. The
BoardEx database contains data on college education, graduate education, past
employment history (including beginning and ending dates of various roles), cur-
rent employment status (including primary employment and outside roles), and
social activities (including memberships, positions held in various foundations
and charitable groups, etc.). The main information we are making use of in this
paper is education, employment history, and demographic information. We col-
lect firm-year patent and citation information from the patent data set created by
Kogan et al. (KPSS) (2017). We collect the inventor information associated with
each patent from the U.S. Patent Inventor Database (1975–2010); see Li et al.
(2014). To calculate control variables, we collect financial statement items from
Compustat and stock price information from the Center for Research in Security
Prices (CRSP). To construct the instrumental variables as we described earlier, we
collect information on mergers and acquisitions from the Securities Data Corpo-
ration (SDC) Mergers & Acquisitions database.
The unique company-level identification code in BoardEx is “Company ID,”
which is unique to BoardEx and cannot be used to merge with other databases
such as Compustat and CRSP. We link the BoardEx database to Compustat and
CRSP in the following way. BoardEx provides the Central Index Key (CIK), the
International Securities Identification Number (ISIN), and the company name.
The “Company ID” in BoardEx is matched with the PERMNO in CRSP by ei-
ther CIK or Committee on Uniform Securities Identification Procedures (CUSIP)
(which is derived from the ISIN). After matching by CIK or CUSIP, we check the
accuracy of the matches by comparing the company names from BoardEx with
the company names from CRSP and Compustat.
The KPSS patent data set provides detailed data for all patents that are
granted by the USPTO over 1926–2011. We use the KPSS patent data rather than
12
One way in which firms with higher-quality management teams may be able to attract higher-
quality inventors is by promoting a more failure-tolerant work environment (in the sense of Manso
(2011)). Manso (2011) has argued that an important variable in encouraging innovation is failure
tolerance. While Manso (2011) does not distinguish between firms with higher- and lower-quality
management teams, if we add the additional assumption that higher-quality managers are also more
failure tolerant, then it will be the case that firms that have higher-quality top management teams will
also have a more failure-tolerant work environment (more conducive to innovative activities).
https://doi.org/10.1017/S0022109018001497
National Bureau of Economic Research (NBER) patent data because the KPSS
patent data enable us to identify comprehensive patent portfolios of firms that
filed patent applications up to 2009, which are granted up to 2011. The NBER
patent data contain patents that have been granted up to 2006, and most of them
have application dates up to 2004. Because our BoardEx sample starts from 1999,
using the KPSS patent data increases our sample size significantly.13 The KPSS
patent data provide the PERMNO for the assignees of each patent. We use this
to merge the patent data with BoardEx as well as Compustat and CRSP. In the
base case analysis, we assign 0 patents to firm-years in the BoardEx sample with-
out any patenting activity. The final BoardEx–KPSS Patent–Compustat–CRSP
merged file leaves us with 6,504 unique firms.
Using the BoardEx employment history file, we identify all the managers
in each matched company for each year from 1999 to 2009. We obtain the sam-
ple of senior managers from BoardEx, whom we define as managers with a ti-
tle of vice president (VP) or higher. The senior managers in our sample can be
broadly categorized into seven groups: CEOs, presidents, chairmen, other chief
officers (chief financial officer (CFO), chief information officer (CIO), etc.), di-
vision heads, VPs, and others. We exclude all firm-years that have at least one
of the following characteristics: i) there is only one manager in the management
team (because it is unlikely that large firms covered by BoardEx have only one
senior manager); ii) there is no CEO for a firm in a given year; iii) there are more
than 30 senior managers in the management team (suggesting that perhaps certain
titles are misleading and we are overclassifying senior managers); iv) financial
and utility firms, defined by Standard Industrial Classification (SIC) codes from
6000 to 6999 and from 4901 to 4999, respectively; and v) firm-years with missing
values for the relevant variables that we need to use. After these exclusions, we
are left with 30,432 firm-year observations for 4,389 firms.
We then obtain the demographic and education information for each senior
manager from the BoardEx database. To obtain education-based connections,
we classify all graduate degrees into four different categories: business school
(MBAs included), medical school, law school, and other graduate (see, e.g.,
Cohen, Frazzini, and Malloy (2008)).
B. Measuring Management Quality
1. Measures of Seven Individual Aspects of Top Management Quality
We measure the quality of a firm’s top management team along two impor-
tant dimensions. The first is based on management team resources, which refers
to the human and knowledge resources (including both education and relevant
work experience) available to top firm management. The second is based on con-
nections available to firm management, which capture their ability to reach out to
managers in other firms, thus enabling them to obtain not only valuable informa-
tion from other firm managers about potential innovation opportunities but also
better terms when dealing with these firms as customers or suppliers, as well as to
hire more and higher-quality inventors.
13
Although BoardEx data start from 1997, data prior to 1999 are sparse (e.g., see Engelberg, Gao,
and Parsons (2013)).
https://doi.org/10.1017/S0022109018001497
The management team resources available to the firm depend in part upon the
number of people in its management team. Therefore, our first measure of team
resources is the size of the firm’s top management team (TEAM SIZE), measured
by the number of officers with the rank of VP or higher. Management team re-
sources also depend upon the knowledge and education of its members. Thus, our
second measure of management team resources is the percentage of the manage-
ment team with an MBA degree (MBA). A larger top management team and a
higher percentage of the team with MBA degrees imply better top management
quality, which will allow the management team to accomplish the tasks discussed
previously. Our third measure of management team resources is the fraction of the
top management team with a Ph.D. degree (PHD). In some innovative firms (e.g.,
technology or biotech firms), some of the top management team may have a Ph.D.
degree, which may help them in choosing the appropriate innovation strategy for
their inventors to work on (even if they themselves are not personally involved
in developing the innovation) as well as in hiring the “right” inventors (scientists
and engineers) who may develop innovations for their firms. Our fourth measure
of management team resources is relevant work experience, which we measure in
two ways. First, we look at the percentage of the management team that has served
as VP or higher in other firms prior to joining the current firm (WORK EXP).
Second, we look at the number of outside board positions that each manager has
previously served on, averaged across all members of the top management team
(BOARD EXP). Clearly, prior board experience may also be a useful asset when
managing a firm because managers may have acquired experience in solving im-
portant problems when serving as board members in the firms for which they have
previously worked.14
We measure top management connections in two ways. First, we look at con-
nections built up by the members of a firm’s top management team based on their
work experience so far (EMP CONN). For each manager, the measure of total
employment-based connections is calculated as the number of senior managers or
directors that each senior manager in the management team has worked with prior
to joining the current firm. If individuals have worked together in the same com-
pany previously during an overlapping time period, they are defined as connected.
In summary, the variable EMP CONN is defined as the number of employment-
based connections of the top management team divided by TEAM SIZE. Second,
we look at the connections built up by members of the firm’s top management
team during their graduate education, which may often last throughout their en-
tire career (EDU CONN). For each manager, the total education-based connec-
tion is calculated as the number of senior managers or directors that each senior
manager in the top management team has been in graduate school with. If indi-
viduals study in the same educational institution, have degrees in the same edu-
cation category (graduate school), and graduate within 1 year of each other, they
are defined as connected. In summary, the variable EDU CONN is defined as the
number of education-based connections of the top management team divided by
TEAM SIZE. In addition, we create the variable AVG TENURE as the average
14
Three of the measures discussed above (TEAM SIZE, MBA, and WORK EXP) are similar to
those used by Chemmanur and Paeglis (2005) and Chemmanur et al. (2011).
https://doi.org/10.1017/S0022109018001497
number of years that each senior manager has worked in a firm, and we use it as
a control variable.
Table 1 provides summary statistics on our seven individual measures of
management team quality. For the median firm in our sample, there are seven se-
nior managers in the management team; 20% of the senior management team has
an MBA degree; 10% of the senior management team has prior work experience
as a senior manager at another firm; 0% of the senior management team of the
median firm has sat on the boards of other firms; and 0% of the senior manage-
ment team has a Ph.D. degree. The median level of EMP CONN is 15.4, and that
of EDU CONN is 0. The median number of years that each manager has worked
in a firm is 5.2 years.
TABLE 1
Summary Statistics
Table 1 reports summary statistics for our sample of public firms between 1999 and 2009. ADJ_PATENT is the truncation-
adjusted number of patents filed by a firm in a given year; ADJ_CITE is the total adjusted number of citations received
by the patents filed by a firm in a given year; ADJ_CPP is the adjusted number of citations per patent filed by a firm
in a given year. TEAM_SIZE is the number of managers (vice president (VP) or higher) in a firm’s management team;
MBA is the fraction of the managers who have MBA degrees; PHD is the fraction of the managers who have Ph.D. de-
grees; WORK_EXP is the fraction of top managers who have experience working as a VP or higher in other companies;
BOARD_EXP is the average number of board positions that each manager has held on; EMP_CONN is the average
number of connections that each manager has through prior employment; EDU_CONN is the average number of grad-
uate connections that each manager has through education (if two managers graduated from the same university with
the same degree within 1 year of each other, those two are defined as connected); TOT_ASSETS is a firm’s total as-
sets; MB is Tobin’s Q, defined as the market value of assets divided by the book value of assets, where the market
value of assets is computed as the book value of assets plus the market value of common stock less the book value
of common stock; ROA is defined as operating income before depreciation divided by total assets; CAPEX/ASSETS is
defined as capital expenditures divided by total assets; R&D/ASSETS is defined as research and development expenses
divided by total assets; RET is a firm’s annual stock return; the Herfindahl–Hirschman index (HHI) for an industry (de-
fined at the 2-digit Standard Industrial Classification (SIC) code level) in a given year is defined by the following formula:
PNo. of firms in the same 2-digit industry
i =1 ((firm salesi )2 /(industry sales)2 ); AVG_TENURE is the average number of years that each
manager has worked as a VP or higher in this firm.
1st 3rd
Variable No. of Obs. Mean Std. Dev. Min Quartile Median Quartile Max
ADJ_PATENT 30,432 0.862 2.308 0.000 0.000 0.000 0.309 10.439
ADJ_CITE 30,432 0.034 0.108 0.000 0.000 0.000 0.000 0.495
ADJ_CPP 30,432 0.002 0.006 0.000 0.000 0.000 0.000 0.025
TEAM_SIZE 30,432 7.751 4.469 2.000 5.000 7.000 10.000 30.000
MBA 30,432 0.230 0.195 0.000 0.000 0.200 0.333 1.000
PHD 30,432 0.074 0.146 0.000 0.000 0.000 0.100 1.000
WORK_EXP 30,432 0.141 0.172 0.000 0.000 0.100 0.235 1.000
BOARD_EXP 30,432 0.071 0.149 0.000 0.000 0.000 0.100 2.667
EMP_CONN 30,432 17.949 11.108 1.000 10.000 15.400 23.000 98.667
EDU_CONN 30,432 1.417 3.296 0.000 0.000 0.000 1.200 55.500
TOT_ASSET ($millions) 30,432 2, 887.979 12,991.060 0.306 83.810 322.609 1, 335.178 304, 594.000
MB 30,432 2.307 2.904 0.166 1.169 1.606 2.499 137.183
ROA 30,432 0.035 0.279 −2.041 0.014 0.102 0.163 0.424
CAPEX/ASSETS 30,432 0.065 0.088 0.000 0.017 0.035 0.074 0.542
R&D/ASSETS 30,432 0.077 0.153 0.000 0.000 0.010 0.092 0.995
RET 30,432 0.173 0.767 −0.884 −0.289 0.033 0.404 3.458
HHI 30,432 0.063 0.060 0.020 0.032 0.041 0.072 1.000
AVG_TENURE 30,432 5.833 3.319 1.000 3.571 5.200 7.333 36.500
2. Common Factor Analysis on Individual Measures of Management Quality

Each of the seven measures described above is likely to have its unique lim-
itations as a measure of the underlying management quality, and is therefore un-
likely to be a comprehensive measure of management quality by itself. Therefore,
we use common factor analysis to capture the variation common to our seven
observable measures of management quality. More precisely, the aim of our fac-
tor analysis is to account for, or explain, the matrix of covariances between our
https://doi.org/10.1017/S0022109018001497
individual measures of management quality using as few factors as possible. Next,

we rotate the initial factors so that each individual measure of management qual-
ity has substantial loadings on as few factors as possible. This methodology is
consistent with the implementation of common factor analysis in the literature.15
All the seven management quality measures above are aggregated to the level
of the management team and are likely to be correlated with firm size. Therefore,
to make sure that these measures are independent of firm size, we use firm size-
and industry-adjusted variables in our common factor analysis. Specifically, we
conduct the following regression for each of the seven proxies of management
quality:
(1) MEASUREi,t = α[ln(FIRM SIZE)i,t ] + β[ln(FIRM SIZE)i,t ]2

+ IND FE + YR FE + i,t ,
where i indexes the firm, and t indexes the year of the observation. Industry
fixed effects defined at the 2-digit SIC code level (IND FE) and year fixed ef-
fects (YR FE) are included. We use the residuals from the above regression as the
firm size- and industry-adjusted measures of the management quality.
Table 2 presents the results of the common factor analysis. The common fac-
tor analysis leads to seven factors. Panel A reports the eigenvalues of each factor.
Factors with higher eigenvalues account for a greater proportion of the variance of
the observed variables. Only the first factor has an eigenvalue that is larger than 1.
Further, we find that this first factor explains 80% of the variation of our individ-
ual management quality proxies (the eigenvalue of the first factor as a proportion
of the sum of the eigenvalues of all seven factors). This suggests that the first fac-
tor is the most important one, providing us with a distinct (unique) measure of
management quality. We term this factor the management quality factor (MQF).16
The first column in Panel B of Table 2 reports the loadings on the first fac-
tor for each of the individual management quality measures. The loadings indicate
that each of the seven individual management quality measures loads positively on
the first factor. Consistent with this, the second column of Panel B shows positive
correlations between the first factor and each of our seven management quality
measures. The last column of Panel B reports the communality of each individual
management quality measure with the common factor, which measures the pro-
portion of the variance of each proxy that is accounted for by the common factor.
15
We adopt common factor analysis rather than principal component analysis as our method of
choice for identifying a single management quality factor. The aim of common factor analysis is to
account for or to “explain” the matrix of covariances between our seven individual management quality
proxies using the minimum number of factors. In contrast, the aim of principal component analysis is
to break down the covariance matrix into a set of orthogonal components equal to the number of the
individual proxies. Given that our objective here is to identify a factor that embodies the underlying
unobservable construct, namely, “management quality,” we believe that the former method is more
appropriate here.
16
In a robustness test that we describe in the Supplementary Material, we address the concern that
our results are driven by the team size measure (TEAM SIZE) alone, but not the other individual
management quality measures. We therefore recalculate the management quality factor by excluding
TEAM SIZE from the common factor analysis. We show that our results are similar when we use the
management quality factor derived from this alternative model.
https://doi.org/10.1017/S0022109018001497
TABLE 2
Extraction of the Management Quality Factor using Common Factor Analysis
Table 2 reports statistics related to our common factor analysis. FACTOR1–FACTOR7 are the common factors obtained by
using common factor analysis on the firm size- and industry-adjusted TEAM_SIZE, MBA, PHD, WORK_EXP, BOARD_EXP,
EMP_CONN, and EDU_CONN. TEAM_SIZE is the number of managers (VP or higher) in a firm’s management team;
MBA is the fraction of the managers who have MBA degrees; PHD is the fraction of the managers have Ph.D. de-
grees; WORK_EXP is the fraction of top managers who have experience working as a VP or higher in other companies;
BOARD_EXP is the average number of board positions that each manager has held; EMP_CONN is the average number
of connections that each manager has through prior employment (if two managers worked in the same previous company
during overlapping time periods, either as managers or directors, those two are defined as connected); EDU_CONN is
the average number of graduate connections that each manager has through education (if two managers graduate from
the same university with the same degree within 1 year of each other, those two are defined as connected). Panel A
reports the eigenvalues for the seven factors to mimic the correlation matrix of the original variables. Panel B reports the
loadings on the first factor and the correlation with the first factor as well as the communality of the original variables.
Panel C reports the descriptive statistics of the first factor.
Panel A. Eigenvalues
FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5 FACTOR 6 FACTOR 7

1.172 0.577 0.216 0.042 −0.022 −0.195 −0.301
Panel B. Summary for Factor Analysis
Variable Loadings on First Factor Correlation with First Factor Communality

TEAM_SIZE 0.513 0.841 0.264
MBA 0.141 0.078 0.020
PHD 0.053 0.038 0.003
WORK_EXP 0.368 0.233 0.135
BOARD_EXP 0.288 0.094 0.083
EMP_CONN 0.803 0.906 0.645
EDU_CONN 0.149 0.086 0.022
Panel C. Summary Statistics of the First Factor
No. of Obs. Mean Std. Dev. Minimum 1st Quartile Median 3rd Quartile Maximum
30,432 0.013 0.758 −1.992 −0.472 −0.040 0.447 2.315
Communality is bounded between 0 and 1, and higher values indicate that a larger
proportion of the variation in the measure is captured by the common factor.
3. Validation of Our Management Quality Factor
This section attempts to validate our management quality factor (MQF) by
using top management team compensation. If MQF indeed captures the true qual-
ity of a firm’s top management team, one immediate implication is that manage-
ment teams with a higher MQF will be paid more than those with a lower MQF.
To test this implication, we make use of the compensation data in the BoardEx
database. BoardEx provides the annual compensation information for a firm’s se-
nior managers, which includes salary, bonus, the value of shares awarded, the
value of the long-term incentive plan (LTIP) awarded, and the value of options
awarded at the manager-firm-year level. However, the coverage of the compen-
sation data is smaller than the coverage of individual measures that are used to
calculate MQF because the compensation information of many managers is miss-
ing in BoardEx. We therefore focus on only the managers with compensation
information available in BoardEx.
For each firm, we construct three different measures of compensa-
tion: AVG TOT COMP, AVG CASH COMP, and EQUITY/TOT COMP. AVG
CASH COMP is defined as the average amount of cash compensation for a
management team, where cash compensation includes base cash salary and
bonus. AVG TOT COMP is defined as the average amount of total compensation
https://doi.org/10.1017/S0022109018001497
for a management team, where total compensation, in addition to cash com-

pensation, also includes the equity compensation, consisting of the value of
shares awarded, the value of LTIP awarded, and the value of options awarded.
EQUITY/TOT COMP is defined as the fraction of equity compensation out of
total compensation. We test the following model:
(2) ln(COMPENSATION)i,t = α + βMQFi,t + γ Z i,t + IND FE
+ YR FE + i,t ,
where ln(COMPENSATION)i,t is the natural logarithm of the previous three com-
pensation measures for the management team in firm i in year t. We take logs
due to the right skewed distribution. Z is a set of control variables, which is de-
scribed in detail in Section III.F. We report the results of the above regressions in
Table 3. Columns 1–3 present regression results with the average total compensa-
tion, the average cash compensation, and the fraction of equity compensation out
of total compensation as dependent variables, respectively. Consistent with our
TABLE 3
The Effect of the Management Quality Factor on Management Team Compensation
Table 3 reports the ordinary least squares (OLS) regression results of various executive compensation measures on our
management quality factor (MQF). For each firm-year, ln(AVG_TOT_COMP) is the natural logarithm of the amount of to-
tal compensation divided by the number of managers. ln(AVG_CASH_COMP) is the natural logarithm of the amount of
cash compensation divided by the number of managers. EQUITY/TOT_COMP is defined as the fraction of equity com-
pensation out of total compensation. Total compensation includes cash compensation and equity compensation. Cash
compensation consists of base cash salary and bonus. Equity compensation consists of the value of shares awarded,
the value of the long-term incentive plan (LTIP) awarded, and the value of options awarded. ln(ASSETS) is the natural
logarithm of a firm’s total assets; MB is Tobin’s Q, defined as the market value of assets divided by the book value of
assets, where the market value of assets is computed as the book value of assets plus the market value of common
stock less the book value of common stock; ROA is defined as operating income before depreciation divided by total
assets; CAPEX/ASSETS is defined as capital expenditures divided by total assets; R&D/ASSETS is defined as research
and development expenses divided by total assets; RET is a firm’s annual stock return; HHI is the industry Herfindahl–
Hirschman index; and AVG_TENURE is the average number of years that each manager has worked as VP or higher in
this firm. Constants, year fixed effects, and 2-digit SIC industry fixed effects are included in all regressions. All standard
errors are adjusted for clustering at the firm level and are reported in parentheses below the coefficient estimates. *, **,
and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
ln(AVG_TOT_COMP) ln(AVG_CASH_COMP) EQUITY/TOT_COMP

Variable 1 2 3
MQF 0.063*** 0.011 0.027***
(0.015) (0.011) (0.004)
ln(ASSETS) 0.472*** 0.254*** 0.064***
(0.009) (0.006) (0.003)
MB 0.478*** 0.119*** 0.090***
(0.029) (0.019) (0.008)
ROA −0.047 0.131** −0.001
(0.075) (0.062) (0.026)
CAPEX/ASSETS 0.049 −0.325** 0.051
(0.210) (0.153) (0.059)
R&D/ASSETS 0.366*** 0.001 0.079*
(0.130) (0.084) (0.045)
RET 0.150*** 0.063*** 0.018***
(0.013) (0.008) (0.004)
HHI 1.035*** −0.613* 0.522***
(0.393) (0.336) (0.162)
AVG_TENURE −0.019*** 0.005 −0.009***
(0.005) (0.004) (0.002)
No. of obs. 12,240 12,232 12,370
Adj. R 2 0.530 0.462 0.193
Industry fixed effects Yes Yes Yes
Year fixed effects Yes Yes Yes
https://doi.org/10.1017/S0022109018001497
expectation, the coefficients on MQF are positive and statistically significant at

the 5% level for two out of three specifications. The economic magnitude of the
effect of MQF is significant as well. For instance, a one interquartile range in-
crease in MQF is associated with a 5.8% increase in the average total compen-
sation. The positive and significant relationship between MQF and management
team compensation suggests that MQF is a valid measure of management quality.
C. Measures of the Quantity and Quality of Corporate Innovation
Following the existing literature (e.g., Kogan et al. (2017), Seru (2014)), we
use patent-based metrics to capture firm innovativeness. Although we also use
R&D expenditures as a measure of investments in innovative activity, patent-
based measures are widely used proxies of innovation output. We obtain patent
data from the database created by KPSS, which provides detailed information on
more than 6 million patents granted by the USPTO from 1926 to 2011. The KPSS
data set has matched assignees in the patent data set with CRSP PERMNOs if the
assignee is a public corporation or subsidiary of a public corporation.
Patent data are subject to two types of truncation problems. First, patents are
recorded in the data set only after they are granted, and the lag between patent
applications and patent grants is significant (approximately 2 years on average).
As we approach the last few years for which patent data are available, we observe
a smaller number of patent applications that are eventually granted. Many patent
applications filed during these years were still under review and had not been
granted by 2011. We partially mitigate this bias by restricting our analyses to 2
years before the patent data end (i.e., in 2009). Further, following Hall, Jaffe, and
Trajtenberg (2001), we correct this bias by dividing each patent for each firm-
year by the mean number of patents for all firms for that year in the same 3-
digit technology class as the patent. The second type of truncation problem stems
from citation counts. Patents tend to receive citations over a long period of time,
so the citation counts of more recent patents are significantly downward biased.
Following Hall et al. (2001) and Seru (2014), this bias is accounted for by scaling
the citations of a given patent by the total number of citations received by all
patents in that year in the same 3-digit technology class as the patent. Note that
this methodology gives us class-adjusted measures of patents and citations, which
adjust for trends in innovative activity in particular industries.
We construct three measures for a firm’s annual innovation output based on
the patent application year.17 The first measure, ln(PATENT), is the natural loga-
rithm of 1 plus the class-adjusted patent count for a firm in a given year. Specifi-
cally, this variable counts the total number of (class-adjusted) patent applications
filed that year that were eventually granted. However, a simple count of patents
may not distinguish breakthrough innovations from incremental technological dis-
coveries. Therefore, we consider two additional measures. The second measure,
ln(CITE), is the natural logarithm of 1 plus the class-adjusted total number of cita-
tions received by a firm’s patents filed in a given year. The third measure, ln(CPP),
17
Consistent with the innovation literature (e.g., Griliches, Pakes, and Hall (1988)), the application
year is more relevant for our purposes than the grant year since it is closer to the time of the actual
innovation.
https://doi.org/10.1017/S0022109018001497
is the natural logarithm of 1 plus the number of citations per patent, constructed
by taking natural logarithm of 1 plus the total number of class-adjusted citations
a firm receives on all the patents it applies for in a given year and normalizing it
by 1 plus the total number of class-adjusted patents applied for in that year. We
take the natural logarithm because the distributions of patents and citations are
right skewed. To avoid losing observations with 0 patents or 0 citations, we add 1
to the actual values. In untabulated analyses, we construct an additional measure
of the quality of corporate innovation using the logged market value of patents
that were filed by a firm in a given year following Kogan et al. (2017), and we
obtain consistent results. Table 1 reports the summary statistics of our innovation
measures. The median R&D to assets ratio in our sample is 1%. Further, an aver-
age (median) firm in our sample has 0.862 (0) class-adjusted patents. An average
(median) firm in our sample has 0.034 (0) class-adjusted citations.
D. Measures of Innovation Strategies
We consider a suite of measures to describe firms’ innovation strategies. The
first set of measures captures the extent to which a firm’s new patents use new
versus existing knowledge. Following Brav et al. (2018), we construct measures
of explorative and exploitative patents (EXPLORE and EXPLOIT), respectively.
EXPLORE (EXPLOIT) is the fraction of explorative (exploitative) patents out of
all the patents filed by a firm in a given year. A patent is explorative if at least 80%
of its citations do not refer to existing knowledge; a patent is exploitative if at least
80% of its citations refer to existing knowledge. Existing knowledge includes a
firm’s previous patent portfolio and all the patents that were cited by the firm’s
patents filed over the past 5 years. Firms venturing into new knowledge and tech-
nologies are likely to generate more explorative innovations, while firms focusing
on existing knowledge and technologies are likely to generate more exploitative
innovations.
Second, following Balsmeier et al. (2017), we categorize our sample pool of
patents into three groups based on their citation distribution among all patents filed
in the same technology class and year: TOP 10, which measures very successful
innovations and is defined as patents receiving the number of citations in the top
10% among all patents in the same 3-digit technology class and application year;
NO CITE, which measures very unsuccessful innovations and is defined as those
receiving 0 citations through the end of our sample period; and M CITE, which
measures moderately successful innovations and is defined as those receiving at
least one citation but not in the top 10%. Firms pursuing innovations in areas that
are newer and less familiar to them are expected to have a greater number of very
successful and very unsuccessful innovations, whereas firms pursuing innovations
that are more familiar to them are expected to have a greater number of moderately
successful innovations.
Third, we consider the number of times a firm’s patents cite patents owned
by the same firm and the number of times a firm’s patents cite patents owned by
other firms (Sørensen and Stuart (2000), Faleye, Hoitash, and Hoitash (2011)).
S CITE is the natural logarithm of 1 plus the number of self-citations (i.e., the
number of times that a firm’s patent portfolio in a given year cites other patents
owned by the same firm). NS CITE is the natural logarithm of 1 plus the number
https://doi.org/10.1017/S0022109018001497
of non-self-citations (i.e., the number of times that a firm’s patent portfolio in

a given year cites patents owned by other firms). Firms pursuing innovations in
areas that are more familiar to them are likely to have more self-citations, whereas
firms pursuing innovations in areas that are newer and less familiar to them are
likely to have more non-self-citations.
Finally, we consider the innovation diversity of a firm’s patent portfolio. We
measure innovation diversity (DIVERSITY or 3YR DIVERSITY) as 1 minus the
Herfindahl index of the number of patents filed by a firm across the 3-digit tech-
nology classes in a given year or in the following 3 years. Firms adopting more
explorative innovation strategies are likely to have greater innovation diversity
(i.e., higher diversification across different technology classes).
E. Measures of Inventor Mobility
To identify inventor mobility, we collect the inventor information of each
patent from the U.S. Patent Inventor Database (1975–2010) (provided by Li et al.
(2014)) from the Harvard Business School Dataverse. The U.S. Patent Inventor
Database includes inventor names, inventor addresses, assignee names, and ap-
plication and grant dates for each patent. More importantly, it identifies unique
inventors over time so that we can potentially track the moves of each inventor.
Following Marx et al. (2009), we identify mobile inventors as changing employ-
ers if they have ever filed two successive patent applications that are assigned to
different entities. Because we need at least two patents to detect a move, inventors
who have filed a single patent throughout their career are necessarily excluded
from our analysis.
For a given firm, an inventor’s move-in year is the year when the inven-
tor filed his or her first patent in this firm; the inventor’s move-out year is that
when the inventor filed his or her first patent in the subsequent firm. For an in-
ventor’s very last employer, we assume that this inventor stayed with that firm
and did not move out.18 For example, in the inventor database, an inventor named
Christopher L. Holderness filed two patent applications through 2010. He filed
patent applications with Corning Inc. in 1999 and then with Dell Inc. in 2003.
Thus, for Corning, Mr. Holderness’s move-in year is 1999, and his move-out year
is 2003, and for Dell, Mr. Holderness’s move-in year is 2003, and he has stayed
with Dell since 2003. Once we identify each mobile inventor’s move-in and move-
out year, we aggregate the number of mobile inventors who move in and move out
at the firm-year level to obtain the total inflows and outflows of mobile inventors
for a given firm in a year. We define the difference between the natural logarithm
of 1 plus the inflow and the natural logarithm of 1 plus the outflow as the net
inflow of mobile inventors (NET INFLOW).
To examine the moves of inventors with different innovative ability, we clas-
sify mobile inventors into two groups, namely, high-quality and low-quality in-
ventors. For each inventor, we look at the average quality of his or her historical
patents (i.e., the citations per patent for all the patents the inventor has filed prior
18
As a robustness check, we redefine the dates that the inventor moved out of the last employer as
1 or 2 years after the inventor filed his or her last patent in that firm. Our results remain qualitatively
similar with this alternative definition.
https://doi.org/10.1017/S0022109018001497
to the current year). If an inventor’s historical citations per patent are higher than
the sample median, the inventor is considered as a high-quality inventor; other-
wise, he or she is a low-quality inventor. We aggregate the number of high-quality
(low-quality) inventors at the firm-year level to get the annual inflow and outflow
of the high-quality (low-quality) inventors for a firm.
F. Other Variables
Following the existing literature, we obtain firms’ financial information from
Compustat and stock price data from CRSP and control for a number of firm char-
acteristics that could affect firms’ innovation output. We compute all variables for
firm i over its fiscal year t. The controls include ln(ASSETS), which is the natural
logarithm of book value of total assets; MB, which is the Tobin’s Q, defined as the
market value of assets divided by the book value of assets, where the market value
of assets is computed as the book value of assets plus the market value of common
stock less the book value of common stock; ROA, which is defined as operating
income before depreciation divided by total assets; CAPEX/ASSETS, which is
defined as capital expenditures over total assets; R&D/ASSETS, which is defined
as research and development expenses divided by total assets; HHI, which is the
value of Herfindahl–Hirschman index in a firm’s industry (defined at the 2-digit
SIC code level) in a given year; RET, which is a firm’s prior-12-month annual
compounded stock return; and AVG TENURE, which is the average number of
years that each manager has worked for a firm. To minimize the effect of outliers,
we winsorize all independent variables at the 1st and 99th percentiles. Table 1
provides summary statistics for the control variables described above. The me-
dian firm size in our sample is $323 million, suggesting that our sample consists
of mainly mid-size and large firms. The median firm in our sample has an ROA of
10.2%, a CAPEX-to-assets ratio of 3.5%, a Tobin’s Q of 1.6, and an annual stock
return of 3.3%.
IV. Empirical Tests and Results

A. The Effect of Management Quality on R&D Expenditures
We empirically test whether firms with higher-quality management teams are
likely to devote a greater amount of resources to innovative activities (Hypothe-
sis 1) by estimating the following regression:
(3) R&D/ASSETSi,t+n = α + βMQFi,t + γ Zi,t + IND FE + YR FE + i,t ,
where i indexes firm; t indexes time; and n equals 1, 2, or 3. Z is a vector of control

variables including ln(ASSETS), MB, ROA, CAPEX/ASSETS, RET, HHI, and
AVG TENURE. We include year fixed effects (YR FE) and 2-digit SIC industry
fixed effects (IND FE).19 In all regressions throughout the paper, standard errors
are clustered at the firm level.20
19
Our results are insensitive to defining industry fixed effects at the 3- or 4-digit SIC code level.
20
We also examine the same regressions controlling for industry, year, and state fixed effects;
industry × year fixed effects; and industry × state × year fixed effects, with similar results. We re-
port the regression results controlling for industry × state × year fixed effects in Table IA-9 in the
Supplementary Material.
https://doi.org/10.1017/S0022109018001497
Table 4 reports the results of equation (3). Columns 1–3 use R&D-to-assets
ratios of 1, 2, and 3 years ahead of the year in which management quality is
measured as dependent variables. We find that the coefficients of MQF in all
three specifications are positive and significant, both statistically and economi-
cally. For instance, the coefficient in column 1 suggests that a one interquartile
range increase in MQF is associated with an increase of 0.74 percentage points
in R&D/ASSETS for next year, which is equivalent to 70% of the sample me-
dian. These results suggest that firms with higher-quality management teams are
associated with greater innovation input, supporting Hypothesis 1.
TABLE 4
The Effect of Management Quality on R&D Expenditures
Table 4 reports the OLS regression results of the ratio of research and development (R&D) expenditures to total assets
on our management quality factor (MQF). Control variables are defined as in Table 3. Constants, year fixed effects, and
2-digit SIC industry fixed effects are included in all regressions. All standard errors are adjusted for clustering at the firm
level and are reported in parentheses below the coefficient estimates. *, **, and *** indicate statistical significance at the
10%, 5%, and 1% levels, respectively.
R&D/ASSETSt +1 R&D/ASSETSt +2 R&D/ASSETSt +3

Variable 1 2 3
MQF 0.008*** 0.008*** 0.007***

(0.001) (0.001) (0.001)
ln(ASSETS) −0.004*** −0.005*** −0.005***
(0.001) (0.001) (0.001)
MB 0.057*** 0.043*** 0.039***
(0.002) (0.002) (0.002)
ROA −0.216*** −0.198*** −0.194***
(0.010) (0.010) (0.010)
CAPEX/ASSETS −0.052*** −0.052*** −0.047***
(0.011) (0.011) (0.012)
RET −0.000 −0.006*** −0.004***
(0.001) (0.001) (0.001)
HHI 0.029* 0.014 0.007
(0.016) (0.017) (0.019)
AVG_TENURE −0.001*** −0.001*** −0.001***
(0.000) (0.000) (0.000)
No. of obs. 27,688 23,741 19,887
Adj. R 2 0.559 0.493 0.472
B. The Effect of Management Quality on Corporate Innovation

Here we study the relation between overall management team quality and
corporate innovation output (Hypothesis 2), both quantity (measured by the num-
ber of patents) and quality (measured by total citations and citations per patent).
We estimate the following model:
(4) INNOV OUTPUTi,t+n = α + βMQFi,t + γ Z i,t

+ IND FE + YR FE + i,t ,
where the dependent variables (INNOV OUTPUT) include ln(PATENT),

ln(CITE), and ln(CPP). Since the innovation process takes time, we examine the
effect of a firm’s management quality on its innovation as of 1, 2, and 3 years after
https://doi.org/10.1017/S0022109018001497
the year in which MQF is measured. Z is the set of control variables similar to
those in the previous section, but it now also includes R&D/ASSETS.21
Table 5 reports the results of the aforementioned tests. Across all the specifi-
cations, the coefficients of MQF are positive and significant, both statistically and
economically.22 For example, column 1 of Panel A suggests that a one interquar-
tile range increase in MQF is associated with a 12.8% increase in the next year’s
adjusted number of patents. These results support Hypothesis 2; that is, firms with
higher-quality management teams are associated with greater and higher-quality
innovation output.23
In untabulated analyses, we further test whether firms with higher-quality
management teams are able to use R&D resources more efficiently in producing
innovation output. Following Hirshleifer et al. (2013), we construct two measures
of innovative efficiency: the number of patents per dollar of R&D expenditures
and the number of citations per dollar of R&D expenditures. Due to space limi-
tations, we report the regression results of innovative efficiency measures on our
management quality factor in Table IA-1 of the Supplementary Material. We find
that our management quality factor is positively and significantly associated with
both measures of innovation efficiency from 1 year up to 3 years from now. Thus,
our evidence indicates that firms with higher-quality management teams are better
at getting more “bang for the buck,” that is, using R&D resources more efficiently
in generating innovation output, supporting Hypothesis 3.
C. Individual Measures of Top Management Quality and Innovation
In this section, we study the relation between various individual aspects
of top management team quality and corporate innovation. Table 6 reports our
regression results of innovation output on each individual management qual-
ity measure. Specifically, we use the values of TEAM SIZE, MBA, PHD,
WORK EXP, BOARD EXP, EMP CONN, and EDU CONN as independent vari-
ables in columns 1–7 across all panels in the table.24 We find that most of our indi-
vidual measures of management quality are positively and significantly associated
with all three innovation output measures. These effects are economically signif-
icant as well. For instance, a 1-interquartile-range increase in MBA is associated
with a 4% increase in the number of patents.
Further, we empirically investigate whether our management quality factor
(MQF) and individual management quality measures have differential effects on
innovation in younger firms (0–2 years after IPO) versus older firms (3 years or
21
In untabulated analyses, we estimate our regressions using the negative binomial maximum-
likelihood estimation technique using the raw number of patents and the total number of citations
received by these patents as dependent variables and obtain qualitatively similar results.
22
In a robustness check, we conduct regressions (reported in Table IA-7 of the Supplementary
Material) using the sample of firms that have filed at least one patent application throughout our
sample period of 1999–2009, and we find similar results.
23
In analyses reported in Table IA-2 of the Supplementary Material, we also use the logged market
value of patents that were filed by a firm in a given year as the dependent variable, following Kogan
et al. (2017), and we obtain consistent results.
24
Our results are similar when we use size-adjusted individual measures of management quality.
https://doi.org/10.1017/S0022109018001497
TABLE 5
The Effect of Management Quality on the Quantity and Quality of Corporate Innovation
Table 5 reports the OLS regression results of the quantity and quality of corporate innovation output on our management
quality factor (MQF). Panels A, B, and C report regression results with the number of patents, the total number of citations,
and the number of citations per patent as dependent variables, respectively. ln(PATENT) is the natural logarithm of 1 plus
the truncation-adjusted number of patents filed by a firm in a given year; ln(CITE) is the natural logarithm of 1 plus the
total adjusted number of citations filed by a firm in a given year; ln(CPP) is the natural logarithm of 1 plus the adjusted
number of citations per patent. Control variables are defined as in Table 3. Constants, year fixed effects, and 2-digit SIC
industry fixed effects are included in all regressions. All standard errors are adjusted for clustering at the firm level and
are reported in parentheses below the coefficient estimates. *, **, and *** indicate statistical significance at the 10%, 5%,
and 1% levels, respectively.
Panel A. The Effect of MQF on the Number of Patents

ln(PATENT)t +1 ln(PATENT)t +2 ln(PATENT)t +3
Variable 1 2 3
MQF 0.128*** 0.129*** 0.128***
(0.011) (0.012) (0.013)
ln(ASSETS) 0.146*** 0.142*** 0.138***
(0.006) (0.006) (0.006)
MB 0.113*** 0.121*** 0.122***
(0.011) (0.012) (0.012)
ROA −0.010 0.002 0.005
(0.021) (0.022) (0.023)
CAPEX/ASSETS −0.043 −0.021 −0.008
(0.064) (0.066) (0.070)
R&D/ASSETS 0.234*** 0.190*** 0.157***
(0.045) (0.045) (0.045)
RET −0.031*** −0.026*** −0.016***
(0.005) (0.005) (0.005)
HHI 0.109 0.067 −0.026
(0.183) (0.191) (0.182)
AVG_TENURE 0.002 0.002 0.002
(0.002) (0.002) (0.002)
No. of obs. 27,688 24,519 21,250
Adj. R 2 0.390 0.384 0.375
Panel B. The Effect of MQF on the Total Number of Citations
ln(CITE)t +1 ln(CITE)t +2 ln(CITE)t +3
Variable 1 2 3
MQF 0.016*** 0.016*** 0.016***

(0.002) (0.002) (0.002)
ln(ASSETS) 0.018*** 0.017*** 0.017***
(0.001) (0.001) (0.001)
MB 0.013*** 0.013*** 0.013***
(0.002) (0.002) (0.002)
ROA −0.010*** −0.010*** −0.008**
(0.003) (0.003) (0.003)
CAPEX/ASSETS −0.006 −0.007 −0.009
(0.010) (0.010) (0.010)
R&D/ASSETS −0.000 −0.005 −0.005
(0.006) (0.006) (0.006)
RET −0.003*** −0.003*** −0.003***
(0.001) (0.001) (0.001)
HHI −0.021 −0.034 −0.035
(0.032) (0.033) (0.031)
AVG_TENURE 0.000 0.000 0.000
(0.000) (0.000) (0.000)
No. of obs. 27,688 24,519 21,250
Adj. R 2 0.256 0.251 0.241
(continued on next page)
https://doi.org/10.1017/S0022109018001497
TABLE 5 (continued)
The Effect of Management Quality on the Quantity and Quality of Corporate Innovation
Panel C. The Effect of MQF on the Number of Citations per Patent
ln(CPP)t +1 ln(CPP)t +2 ln(CPP)t +3

Variable 1 2 3
MQF 0.001*** 0.001*** 0.001***
(0.000) (0.000) (0.000)
ln(ASSETS) 0.001*** 0.001*** 0.001***
(0.000) (0.000) (0.000)
MB 0.000*** 0.001*** 0.000***
(0.000) (0.000) (0.000)
ROA −0.000 −0.000 −0.000
(0.000) (0.000) (0.000)
CAPEX/ASSETS −0.000 −0.000 −0.001
(0.000) (0.000) (0.001)
R&D/ASSETS 0.001** 0.000 0.000
(0.000) (0.000) (0.000)
RET −0.000* −0.000 −0.000*
(0.000) (0.000) (0.000)
HHI −0.005** −0.005** −0.004
(0.002) (0.002) (0.002)
AVG_TENURE 0.000 0.000 0.000
(0.000) (0.000) (0.000)
No. of obs. 27,688 24,519 21,250
Adj. R 2 0.140 0.139 0.136
greater after IPO).25 We perform our baseline regressions (as in equation (4)) in
younger and older firms and test whether the effects of MQF and each individual
measure on innovation are significantly different across these two kinds of firms.
We report the results of these tests in Table 7.26 In columns 1 and 2, we first ana-
lyze the relation between MQF and corporate innovation in younger versus older
firms. We find that MQF is positively and significantly associated with the num-
ber of patents for both groups and that the difference of the effects of MQF on
the number of patents is not significant. When we analyze individual measures
of management quality, we observe some interesting heterogeneity in the relation
between these individual measures and corporate innovation for younger firms
versus older firms. We find that management team size, the fraction of managers
with MBA degrees, and education-based connections positively and significantly
affect corporate innovation only for older firms.27 Prior managerial work expe-
rience positively and significantly affects corporate innovation only for younger
firms. Two individual measures, namely, the fraction of top managers with Ph.D.
degrees and employment-based connections, positively and significantly affect
corporate innovation in both younger and older firms; however, the magnitude
25
We collect the IPO date for each firm from the SDC Platinum New Issues Database. We cross-
check the IPO date with Compustat and drop observations for which the two data sources provide
inconsistent IPO dates.
26
We obtain consistent results using citations and citations per patent across different time horizons
as dependent variables.
27
Note that the differences in the regression coefficients on education-based connections are not
significantly different for younger versus older firms.
https://doi.org/10.1017/S0022109018001497
TABLE 6
The Effect of Individual Management Quality Measures on Corporate Innovation
Table 6 reports the OLS regression results of corporate innovation on individual management quality measures. Panels A,
B, and C report regression results with the number of patents, the total number of citations, and the number of citations
per patent as dependent variables, respectively. ln(PATENT) is the natural logarithm of 1 plus the truncation-adjusted
number of patents filed by a firm in a given year; ln(CITE) is the natural logarithm of 1 plus the total adjusted number
of citations filed by a firm in a given year; ln(CPP) is the natural logarithm of 1 plus the adjusted number of citations
per patent. Individual management quality measures are defined in detail in Table 2. Control variables are the same as
in Table 3 in all regressions and coefficient estimates on controls are not reported to save space. Constants, year fixed
effects, and 2-digit SIC industry fixed effects are included in all regressions. All standard errors are adjusted for clustering
at the firm level and are reported in parentheses below the coefficient estimates. The coefficients and standard errors in
Panel C are multiplied by 100 for ease of reading. *, **, and *** indicate statistical significance at the 10%, 5%, and 1%
levels, respectively.
Panel A. The Effect of Individual Management Quality Measures on the Number of Patents
ln(PATENT)t +1
Variable 1 2 3 4 5 6 7
TEAM_SIZE 0.029***
(0.003)
MBA 0.120***
(0.033)
PHD 0.356***
(0.057)
WORK_EXP 0.035
(0.035)
BOARD_EXP −0.002
(0.026)
EMP_CONN 0.013***
(0.001)
EDU_CONN 0.012***
(0.002)
No. of obs. 27,688 27,688 27,688 27,688 27,688 27,688 27,688
Adj. R 2 0.389 0.370 0.374 0.369 0.369 0.391 0.372
Controls Yes Yes Yes Yes Yes Yes Yes
Industry fixed effects Yes Yes Yes Yes Yes Yes Yes
Year fixed effects Yes Yes Yes Yes Yes Yes Yes
Panel B. The Effect of Individual Management Quality Measures on the Total Number of Citations
ln(CITE)t +1
Variable 1 2 3 4 5 6 7
TEAM_SIZE 0.004***
(0.000)
MBA 0.013***
(0.005)
PHD 0.021***
(0.007)
WORK_EXP −0.006
(0.005)
BOARD_EXP −0.002
(0.003)
EMP_CONN 0.002***
(0.000)
EDU_CONN 0.001***
(0.000)
No. of obs. 27,688 27,688 27,688 27,688 27,688 27,688 27,688
Adj. R 2 0.261 0.240 0.240 0.239 0.239 0.259 0.242
(continued on next page)
https://doi.org/10.1017/S0022109018001497
TABLE 6 (continued)
The Effect of Individual Management Quality Measures on Corporate Innovation
Panel C. The Effect of Individual Management Quality Measures on the Number of Citations per Patent
ln(CPP)t +1
Variable 1 2 3 4 5 6 7
TEAM_SIZE 0.013***
(0.002)
MBA 0.059**
(0.027)
PHD 0.112***
(0.042)
WORK_EXP −0.004
(0.028)
BOARD_EXP −0.024
(0.018)
EMP_CONN 0.005***
(0.001)
EDU_CONN 0.005***
(0.002)
No. of obs. 27,688 27,688 27,688 27,688 27,688 27,688 27,688
Adj. R 2 0.140 0.136 0.136 0.135 0.135 0.139 0.136
of the effect of the Ph.D. measure is greater for corporate innovation in younger
firms. These results are broadly consistent with Hypothesis 4.
D. Management Quality and Corporate Innovation Strategies
In this section, we investigate the effect of management quality on firms’
innovation strategies. As we describe in Section III.D, we adopt a suite of
measures to capture the extent to which firms engage in explorative versus ex-
ploitative innovation strategies.
First, we study the relation between management quality and the fraction of
explorative patents (using at least 80% new knowledge) and that of exploitative
patents (using at least 80% existing knowledge) out of all the patents filed by a
firm in a given year (i.e., EXPLORE and EXPLOIT, respectively). Table 8 re-
ports the results of these tests. We find that the coefficients of MQF are positive
and significant across all specifications, suggesting that firms with higher man-
agement quality are associated with a greater fraction of explorative patents as
well as a greater fraction of exploitative patents. Further, we find that the coeffi-
cients on MQF in the odd columns are significantly greater than those in the even
columns, suggesting that firms with higher-quality management teams are likely
to engage more in explorative innovation strategies relative to exploitative innova-
tion strategies, thus producing more innovations venturing into the development
of new knowledge and technologies.
We use several other measures to describe a firm’s innovation strategy and
report these results in Table 9. In Panel A, following Balsmeier et al. (2017),
we study the relation between management quality and a firm’s innovation strat-
egy by looking at the very successful (TOP 10), very unsuccessful (NO CITE),
TABLE 7
The Effect of Management Quality Factor and Individual Measures of Management Quality on Innovation for Younger versus Older Firms
Table 7 reports the OLS regression results of corporate innovation on our management quality factor (MQF) and individual measures of management quality for younger (0–2 years after initial public offering
(IPO)) versus older firms (3 years or greater after IPO). ln(PATENT) is the natural logarithm of 1 plus the truncation-adjusted number of patents filed by a firm in a given year. Individual management quality
measures are defined in detail in Table 2. Control variables are the same as in Table 3 and coefficients are not reported to save space. Constants, year fixed effects and 2-digit SIC industry fixed effects are
included in all regressions. All standard errors are adjusted for clustering at the firm level and are reported in parentheses below the coefficient estimates. *, **, and *** indicate statistical significance at the
Panel A. The Effect of Management Quality Measures on the Number of Patents for Younger Firms versus Older Firms
ln(PATENT)t +1
Younger Older Younger Older Younger Older Younger Older Younger Older Younger Older Younger Older Younger Older
Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
MQF 0.067** 0.084***

(0.027) (0.017)
TEAM_SIZE 0.006 0.018***
(0.007) (0.004)
MBA −0.049 0.064***
Chemmanur, Kong, Krishnan, and Yu

(0.049) (0.022)
PHD 0.659*** 0.490***
(0.135) (0.082)
WORK_EXP 0.249*** 0.059
(0.091) (0.052)
BOARD_EXP 0.096 −0.013
(0.095) (0.041)
EMP_CONN 0.009*** 0.008***
(0.003) (0.002)
EDU_CONN 0.008 0.008***
(0.005) (0.003)
No. of obs. 1,923 11,212 1,923 11,212 1,923 11,212 1,923 11,212 1,923 11,212 1,923 11,212 1,923 11,212 1,923 11,212
Adj. R 2 0.285 0.342 0.281 0.340 0.281 0.332 0.315 0.346 0.287 0.332 0.281 0.332 0.289 0.340 0.282 0.335
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Industry fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Panel B. The Differences of the Effects of Management Quality Measures on the Number of Patents
Difference Difference Difference Difference Difference Difference Difference Difference

1–2 3–4 5–6 7–8 9–10 11–12 13–14 15–16
27
Difference −0.017 −0.012* −0.113** 0.169* 0.190** 0.110 0.000 −0.001
https://doi.org/10.1017/S0022109018001497
https://doi.org/10.1017/S0022109018001497
TABLE 8
The Effect of Management Quality on Explorative versus Exploitative Innovations
Panel A of Table 8 reports the OLS regression results of a firm’s explorative innovations and exploitative innovations on
our management quality factor (MQF). EXPLORE is the fraction of explorative patents out of all the patents filed by a firm
in a given year; EXPLOIT is the fraction of exploitative patents out of all the patents filed by a firm in a given year. A patent
is explorative if at least 80% of its citations do not refer to existing knowledge, which includes a firm’s previous patent
portfolio and all the patents that were cited by a firm’s patents filed over the past 5 years. A patent is exploitative if at
least 80% of its citations refer to existing knowledge, which includes a firm’s previous patent portfolio and all the patents
that have been cited by a firm’s patents filed over the past 5 years. Panel B reports the difference between the coefficient
estimates of MQF using EXPLORE and EXPLOIT as dependent variables and tests their statistical differences. Control
variables are defined as in Table 3. Constants, year fixed effects and 2-digit SIC industry fixed effects are included in all
regressions. All standard errors are adjusted for clustering at the firm level and are reported in parentheses below the
coefficient estimates. *, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
Panel A. The Effect of MQF on Explorative and Exploitative Patents
EXPLOREt +1 EXPLOITt +1 EXPLOREt +2 EXPLOITt +2 EXPLOREt +3 EXPLOITt +3

Variable 1 2 3 4 5 6
MQF 0.027*** 0.016*** 0.026*** 0.016*** 0.028*** 0.016***
(0.004) (0.002) (0.004) (0.002) (0.004) (0.002)
ln(ASSETS) 0.037*** 0.013*** 0.036*** 0.013*** 0.034*** 0.014***
(0.002) (0.001) (0.002) (0.001) (0.002) (0.001)
MB 0.025*** 0.012*** 0.025*** 0.014*** 0.024*** 0.016***
(0.005) (0.002) (0.005) (0.002) (0.005) (0.003)
ROA 0.022** −0.001 0.015 0.005 0.016 0.008
(0.011) (0.006) (0.011) (0.006) (0.011) (0.006)
CAPEX/ASSETS 0.007 −0.017 0.034 −0.010 0.021 0.002
(0.026) (0.013) (0.027) (0.014) (0.028) (0.015)
R&D/ASSETS 0.225*** 0.074*** 0.168*** 0.077*** 0.140*** 0.095***
(0.022) (0.012) (0.021) (0.012) (0.021) (0.014)
RET −0.002 −0.002 −0.005 −0.004*** 0.004 −0.002
(0.003) (0.001) (0.003) (0.001) (0.003) (0.001)
HHI 0.177* 0.051 0.114 0.054 −0.050 0.078*
(0.096) (0.036) (0.100) (0.038) (0.130) (0.041)
AVG_TENURE −0.001 0.001* −0.000 0.001 0.000 0.000
(0.001) (0.000) (0.001) (0.000) (0.001) (0.000)
No. of obs. 27,688 27,688 24,519 24,519 21,250 21,250
Adj. R 2 0.214 0.117 0.207 0.116 0.202 0.121
Industry fixed effects Yes Yes Yes Yes Yes Yes
Year fixed effects Yes Yes Yes Yes Yes Yes
Panel B. The Differences of the Effects of MQF on Explorative versus on Exploitative Patents
Difference Difference Difference

1–2 3–4 5–6
Difference 0.011*** 0.010*** 0.013***
and moderately successful innovations (M CITE), based on their citation distri-

bution among all patents filed in the same technology class and year. We find that
management quality is positively and significantly associated with the number of
patents in all three categories. More interestingly, management quality has a more
pronounced effect on successful patents than on unsuccessful patents or average
patents; that is, firms with higher-quality management teams are better at motivat-
ing the development of patents that are highly cited afterward.28 In Panel B, we
study the relation between management quality and non-self-citations (i.e., the
number of times a firm’s patents cite other firms’ patents) and self-citations (i.e.,
the number of times a firm’s patents cite its own patents). We also test whether the
effect of management quality on non-self-citations is statistically different from
28
We obtain qualitatively similar results using all three categories of innovations across other time
horizons as dependent variables (Table IA-3 of our Supplementary Material).
https://doi.org/10.1017/S0022109018001497
TABLE 9
The Effect of Management Quality on Other Measures of Innovation Strategies
Panel A of Table 9 reports the OLS regression results of the highly successful (TOP_10), unsuccessful (NO_CITE) and
moderately successful innovations (M_CITE) on our management quality factor (MQF). Panel B reports and compares
the regression results of self-citations and non-self-citations on MQF. Panel C reports the regression results of innovation
diversity on MQF. TOP_10 is the natural logarithm of 1 plus a firm’s number of patents that received cites within the top 10%
among all patents in the same 3-digit patent class and application year; NO_CITE is the natural logarithm of 1 plus the
number of patents that received no citation; M_CITE is the natural logarithm of 1 plus the number of patents that received
at least 1 citation but below the top 10% among all patents. NS_CITE is the natural logarithm of 1 plus the number of
non-self-citations. S_CITE is the natural logarithm of 1 plus the number of self-citations. DIVERSITY (3YR_DIVERSITY) is 1
minus the Herfindahl index of the number of patents filed by a firm across the 3-digit technology classes in a given year (in
the following 3 years). Control variables are the same as in Table 3 in all regressions and coefficients are not reported to
save space. Constants, year fixed effects and 2-digit SIC industry fixed effects are included in all regressions. All standard
Panel A. The Effect of MQF on the Highly Successful, Unsuccessful, and Moderately Successful Innovations

TOP_10t +1 NO_CITEt +1 M_CITEt +1 1–2 1–3 2–3
MQF 0.303*** 0.147*** 0.144*** 0.156*** 0.159*** 0.003

(0.038) (0.018) (0.020) (0.023) (0.022) (0.008)
No. of obs. 15,251 15,251 15,251
Adj. R 2 0.406 0.384 0.428
Controls, industry and year Yes Yes Yes
fixed effects
Panel B. The Effect of MQF on Non-Self-Citations and Self-Citations
NS_CITEt +1 S_CITEt +1 NS_CITEt +2 S_CITEt +2 NS_CITEt +3 S_CITEt +3

MQF 0.316*** 0.253*** 0.313*** 0.251*** 0.304*** 0.244***

(0.028) (0.023) (0.030) (0.024) (0.031) (0.026)
No. of obs. 27,688 27,688 24,519 24,519 21,250 21,250
Adj. R 2 0.389 0.331 0.382 0.328 0.374 0.322
Controls, industry and year Yes Yes Yes Yes Yes Yes
fixed effects
1–2 3–4 5–6
Difference 0.062*** 0.062*** 0.060***

Panel C. The Effect of MQF on Innovation Diversity
DIVERSITYt +1 DIVERSITYt +2 DIVERSITYt +3 3YR_DIVERSITYt +1

Variable 1 2 3 4
MQF 0.036*** 0.039*** 0.036*** 0.041***

(0.006) (0.007) (0.007) (0.007)
No. of obs. 8,295 7,016 5,750 8,851
Adj. R 2 0.355 0.359 0.358 0.328
Controls, industry and year Yes Yes Yes Yes
fixed effects
that on self-citations and report these test results at the bottom of Panel B. We
find that firms with higher-quality management teams are associated with a sig-
nificantly greater number of non-self-citations than self-citations, suggesting that
such firms tend to pursue innovations in areas that are newer and less familiar
to them. In Panel C, we study the relation between management quality and in-
novation diversity, which captures the level of diversification of a firm’s patent
portfolio over different technology classes. We find that firms with higher-quality
management teams are associated with a greater level of innovation diversity for
the following 3 years. Collectively, our empirical results presented in this section
support the argument that firms with higher-quality management teams are more
https://doi.org/10.1017/S0022109018001497
likely to engage in explorative innovation strategies, those involving new tech-

nologies, and those that are likely to push the knowledge boundaries of the firm
outward (Hypothesis 5A).
E. Management Quality and Corporate Innovation across Industries
In untabulated analyses, we investigate whether the positive and significant
relationship established earlier between our MQF and corporate innovation is
stronger in some industries than in others. We find that this relationship is stronger
in financially constrained industries (as measured by the level of external financial
dependence; see, e.g., Rajan and Zingales (1998)) and in more competitive indus-
tries (as measured by the Herfindahl–Hirschman index in the firm’s industry).
Due to space limitations, we report these results in Table IA-4 of our Supplemen-
tary Material. These results suggest that firms with higher-quality management
teams are able to select better projects, use resources more efficiently, and gen-
erate greater innovation output in adverse financing environments and in more
competitive industries. These empirical findings lend support for Hypothesis 6
and Hypothesis 7.
V. Analysis of the Inventor Mobility Channel

Our evidence thus far shows that top management quality is positively related
to corporate innovation. In this section, we analyze an important channel through
which this may occur. As argued before, higher-quality management teams may
provide more R&D resources, manage R&D resources better, and provide a more
risk-tolerant climate for inventors to succeed in. This, in turn, may make firms
with higher-quality management teams more attractive to higher-quality inven-
tors. Thus, one way that higher-quality management teams may enhance innova-
tion is by hiring more and higher-quality inventors to work for the firm. In this
section, we test these hypotheses.
A. Management Quality and the Net Inflow of Inventors
To assess the relation between management quality and the net inflow of in-
ventors who move into a firm in a given year (Hypothesis 8), we test the following
model:
(5) NET INFLOWi,t+n = α + βMQFi,t + γ Zi,t + IND FE

+ YR FE + ST FE + i,t ,
where i indexes firm; t indexes time; and n equals 1, 2, or 3. Z is a vector of

control variables used in prior tests. As before, we include year fixed effects and
2-digit SIC industry fixed effects. Further, since location may impact an inventor’s
decision of moving into or out of a firm, we include state fixed effects (ST FE)
for the state of the firm’s headquarters in all regressions in this section.
Table 10 reports the results of these tests. The coefficients of MQF are pos-
itive and statistically and economically significant across all specifications. For
instance, column 1 suggests that a one interquartile range increase in MQF is as-
sociated with a 0.05 increase in NET INFLOW. The economic magnitude of such
an effect is significant, given that the sample mean of NET INFLOW is 0.21.
https://doi.org/10.1017/S0022109018001497
TABLE 10
The Effect of Management Quality on the Net Inflow of Inventors
Table 10 reports the OLS regression results of the net inflow of inventors for a firm in a given year on our management
quality factor (MQF). NET_INFLOW is defined as the difference between the inflow and outflow of inventors, where the
inflow and outflow of inventors are defined as the natural logarithm of 1 plus the total number of inventors who move in
and who move out aggregated at the firm-year level, respectively. Control variables are defined as in Table 3. Constants;
year fixed effects; 2-digit SIC industry fixed effects; and state fixed effects are included in all regressions. All standard
NET_INFLOWt +1 NET_INFLOWt +2 NET_INFLOWt +3

Variable 1 2 3
MQF 0.050*** 0.049*** 0.054***
(0.007) (0.007) (0.007)
ln(ASSETS) 0.073*** 0.067*** 0.060***
(0.003) (0.003) (0.003)
MB 0.076*** 0.080*** 0.076***
(0.008) (0.008) (0.008)
ROA 0.051*** 0.053*** 0.050***
(0.016) (0.015) (0.016)
CAPEX/ASSETS 0.123** 0.104** 0.070
(0.050) (0.048) (0.050)
R&D/ASSETS 0.356*** 0.257*** 0.196***
(0.039) (0.035) (0.036)
RET −0.025*** −0.016*** −0.007*
(0.005) (0.004) (0.004)
HHI 0.344** 0.273* 0.136
(0.146) (0.140) (0.156)
AVG_TENURE −0.004*** −0.003*** −0.002*
(0.001) (0.001) (0.001)
No. of obs. 25,945 23,096 20,119
Adj. R 2 0.255 0.244 0.234
Industry, year, and state fixed effects Yes Yes Yes
These findings support Hypothesis 8; that is, an important channel through which
higher-quality management teams enhance corporate innovation is by hiring more
inventors.
In an untabulated analysis, we include the net inflow of inventors in the ordi-
nary least squares (OLS) regressions for a firm’s innovation output. Therefore, we
regress our corporate innovation output measures on the net inflow of inventors
and our MQF while including the same of set of control variables as in Table 5,
year fixed effects, industry fixed effects, and state fixed effects. This test allows
us to check whether MQF still has a direct (residual) effect on innovation after
controlling for the net inflow of inventors. We find that whereas the coefficients
on the net inflow of inventors all remain positive and significant, the coefficients
on MQF are still positive and significant at the 1% level. However, the magni-
tudes of the coefficients on MQF are reduced quite substantially once we include
the net inflow of inventors (e.g., the coefficient on MQF is decreased by 28%
when ln(PATENT)t+1 is the dependent variable, and the decrease is statistically
significant at the 1% level), suggesting that the effect of management quality on
innovation is at least partly channeled through the net inflow of inventors.
B. Management Quality and High- and Low-Quality Inventors
In this section, we move on to test whether firms with higher-quality manage-
ment teams are better at attracting higher-quality inventors. We therefore conduct
https://doi.org/10.1017/S0022109018001497
regressions as in equation (5), using the net inflow of high-quality inventors and
that of low-quality inventors as dependent variables, respectively, and test whether
the coefficients of our management quality factor in these regressions are signifi-
cantly different from each other.
Panel A of Table 11 reports the regression results using
NET INFLOW HIGHi,t and NET INFLOW LOWi,t as dependent variables
calculated at 1, 2, and 3 years subsequent to the current year. We find that the
coefficients on MQF are positive using both dependent variables, suggesting that
management quality has positive impacts on the net inflow of both high-quality
and low-quality inventors. More importantly, the effect of MQF on the net
inflow of high-quality inventors is economically 10 times larger than on the net
inflow of low-quality inventors across all time horizons. We test the statistical
TABLE 11
The Effect of Management Quality on the Net Inflow of High-Quality and Low-Quality
Inventors
Panel A of Table 11 reports the OLS regression results of the net inflow of high-quality inventors and the net inflow of low-
quality inventors for a firm in a given year on our management quality factor (MQF). NET_INFLOW_HIGH is the difference
between the natural logarithm of 1 plus the number of high-quality inventors who move into the firm and the natural
logarithm of 1 plus the number of high-quality inventors who move out of the firm in a given year; NET_INFLOW_LOW
is the difference between the natural logarithm of 1 plus the number of low-quality inventors who move into the firm and
the natural logarithm of 1 plus the number of low-quality inventors who move out of the firm in a given year. Panel B
reports the difference between the coefficient estimates of MQF using NET_INFLOW_HIGH and NET_INFLOW_LOW as
dependent variables and tests their statistical differences. Control variables are defined as in Table 3. Constants; year
fixed effects; 2-digit SIC industry fixed effects; and state fixed effects are included in all regressions. All standard errors
are adjusted for clustering at the firm level and are reported in parentheses below the coefficient estimates. *, **, and ***
indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
Panel A. The Effects of MQF on the Net Inflow of High-Quality and Low-Quality Inventors
NET_INFLOW_ NET_INFLOW_ NET_INFLOW_ NET_INFLOW_ NET_INFLOW_ NET_INFLOW_

HIGHt +1 LOWt +1 HIGHt +2 LOWt +2 HIGHt +3 LOWt +3
MQF 0.054*** 0.004*** 0.053*** 0.004*** 0.056*** 0.004***
(0.007) (0.001) (0.007) (0.001) (0.007) (0.001)
ln(ASSETS) 0.078*** 0.004*** 0.071*** 0.004*** 0.064*** 0.003***
(0.003) (0.000) (0.003) (0.000) (0.003) (0.000)
MB 0.079*** 0.002*** 0.084*** 0.003*** 0.080*** 0.003***
(0.009) (0.001) (0.009) (0.001) (0.008) (0.001)
ROA 0.050*** 0.001 0.053*** 0.001 0.051*** 0.002*
(0.016) (0.001) (0.016) (0.001) (0.016) (0.001)
CAPEX/ASSETS 0.123** 0.004 0.112** 0.003 0.082 −0.003
(0.051) (0.004) (0.050) (0.004) (0.051) (0.004)
R&D/ASSETS 0.356*** 0.013*** 0.264*** 0.008*** 0.206*** 0.008**
(0.040) (0.004) (0.036) (0.003) (0.038) (0.003)
RET −0.025*** −0.001** −0.018*** −0.000 −0.008* −0.001**
(0.005) (0.000) (0.005) (0.000) (0.004) (0.000)
HHI 0.393*** −0.001 0.313** 0.001 0.161 0.001
(0.151) (0.011) (0.145) (0.010) (0.162) (0.010)
AVG_TENURE −0.004*** 0.000 −0.003*** 0.000 −0.002* 0.000
(0.001) (0.000) (0.001) (0.000) (0.001) (0.000)
No. of obs. 25,945 25,945 23,096 23,096 20,119 20,119
Adj. R 2 0.263 0.063 0.253 0.061 0.241 0.060
Industry, year, and Yes Yes Yes Yes Yes Yes
state fixed effects
Panel B. Differences of the Effects of MQF on the Net Inflow of High-Quality and Low-Quality Inventors

1–2 3–4 5–6
Difference 0.050*** 0.049*** 0.052***

https://doi.org/10.1017/S0022109018001497
significance of the difference between the coefficients on MQF for high-quality

versus low-quality inventors and report these test results in Panel B. We find that
all these differences are statistically significant at the 1% level.
In untabulated analyses, we also investigate whether management quality
is positively and significantly associated with the change in the average inven-
tor quality of a firm. The quality of each inventor is measured as the citations
per patent for the patents the inventor has filed prior to the current year. The net
change in the average inventor quality of a firm in a given year is defined as the
difference between the average quality of all incoming inventors joining the firm
and that of all outgoing inventors leaving the firm in a given year. We find that our
management quality factor is indeed positively and significantly associated with
the net change in the average inventor quality.29 Collectively, our empirical find-
ings provide strong evidence that firms with higher-quality management teams are
able to hire a greater number of high-quality inventors than low-quality inventors,
consistent with Hypothesis 9.
To further analyze whether hiring high-quality inventors is indeed a chan-
nel through which firms with higher-quality management teams spur innovation,
we include the net inflow of high-quality inventors in our OLS regressions for
innovation output. If the relation between management quality and innovation is
channeled through the hiring of high-quality inventors, we would expect the co-
efficients of the net inflow of high-quality inventors to be positive and significant,
whereas the magnitudes of the coefficients on MQF should become significantly
lower.
Table 12 reports the results of these regressions. Panels A, B, and C use the
number of patents, the total number of citations, and citations per patent, respec-
tively, as dependent variables. The same set of control variables and fixed effects
as in Table 11 is included in all the regressions in Table 12. Consistent with our
hypothesis, we find that the coefficients of the net inflow of high-quality inventors
are positive and significant at the 1% level in all the regressions in these three pan-
els. Further, the coefficients of our management quality factor (MQF) have much
smaller magnitudes compared to the coefficients from regressions in which the net
inflow of higher-quality inventors is not included (i.e., compared to our Table 5
results), suggesting that the effect of MQF on corporate innovation is partially
mediated through hiring higher-quality inventors. For instance, the coefficient on
MQF with the subsequent year adjusted patents as the dependent variable is 0.096
when we control for the inflow of high-quality inventors, whereas it is 0.146 with-
out this control (in Panel A of Table 5), reflecting a 34% decline. More interest-
ingly, we find that the coefficients of MQF are smaller in the regressions where
the net inflow of high-quality inventors is included compared to the case where
the net inflow of all inventors is included.30 Collectively, these results support our
conjecture that an important channel through which higher-quality management
teams enhance innovation is by hiring higher-quality inventors.
29
Due to space limitations, we report these results in Table IA-5 of our Supplementary Material.
30
In untabulated analysis, we find that the differences between the coefficients on MQF in these
two sets of regressions are statistically significant.
https://doi.org/10.1017/S0022109018001497
TABLE 12
The Effect of the Net Inflow of High-Quality Inventors on Corporate Innovation
Table 12 reports the OLS regression results of corporate innovation on the net inflow of high-quality inventors and our
management quality factor (MQF). Panels A, B, and C report regression results with the number of patents, the total
number of citations, and the number of citations per patent as dependent variables, respectively. ln(PATENT) is the
natural logarithm of 1 plus the truncation-adjusted number of patents filed by a firm in a given year; ln(CITE) is the natural
logarithm of 1 plus the total adjusted number of citations filed by a firm in a given year; ln(CPP) is the natural logarithm of
1 plus the adjusted number of citations per patent. NET_INFLOW_HIGH is the difference between the natural logarithm
of 1 plus the number of high-quality inventors who move into the firm and the natural logarithm of 1 plus the number
of high-quality inventors who move out of the firm in a given year. Control variables are the same as in Table 3 in all
regressions and coefficients are not reported to save space. Constants; year fixed effects; 2-digit SIC industry fixed
effects; and state fixed effects are included in all regressions. All standard errors are adjusted for clustering at the firm
level and are reported in parentheses below the coefficient estimates. *, **, and *** indicate statistical significance at the
Panel A. MQF, the Net Inflow of High-Quality Inventors, and the Number of Patents
ln(PATENT)t +1 ln(PATENT)t +2 ln(PATENT)t +3

Variable 1 2 3
NET_INFLOW_HIGH 0.572*** 0.526*** 0.466***
(0.013) (0.014) (0.014)
MQF 0.096*** 0.093*** 0.090***
(0.008) (0.009) (0.010)
No. of obs. 25,945 23,096 20,119
Adj. R 2 0.578 0.559 0.528
Controls, industry, year, and state fixed effects Yes Yes Yes
Panel B. MQF, the Net Inflow of High-Quality Inventors, and the Total Number of Citations
ln(CITE)t +1 ln(CITE)t +2 ln(CITE)t +3

Variable 1 2 3
NET_INFLOW_HIGH 0.063*** 0.058*** 0.051***
(0.003) (0.003) (0.003)
MQF 0.014*** 0.013*** 0.013***
(0.001) (0.002) (0.002)
No. of obs. 25,945 23,096 20,119
Adj. R 2 0.372 0.360 0.334
Panel C. MQF, the Net Inflow of High-Quality Inventors, and the Number of Citations per Patent
ln(CPP)t +1 ln(CPP)t +2 ln(CPP)t +3

Variable 1 2 3
NET_INFLOW_HIGH 0.002*** 0.002*** 0.002***
(0.000) (0.000) (0.000)
MQF 0.000*** 0.000*** 0.000***
(0.000) (0.000) (0.000)
No. of obs. 25,945 23,096 20,119
Adj. R 2 0.181 0.184 0.184
VI. Instrumental Variable Analysis

In this section, we address the potential endogeneity concern discussed in
the Introduction using an IV analysis. The instrument we use is the number of
acquisitions in the industry and state of the sample firm 5 years prior weighted
by an index measuring the enforceability index of noncompete clauses in that
state aggregated to the national level. Our instrument is motivated by the follow-
ing facts. First, potential managers available for hire by a firm often come from
established firms in the same industry and may leave such firms as a result of ac-
quisitions. In other words, there is a strong correlation between the movement of
managers across firms and the number of acquisitions in the industry that the firm
https://doi.org/10.1017/S0022109018001497
belongs to.31 Second, the enforceability of noncompete clauses, which are com-
monly used in the employment contracts for top management teams to prohibit
them from joining or founding a rival company within 1 or 2 years of leaving,
affects the mobility of managers across firms.32,33,34
Specifically, the instrumental variable for the management quality factor
(MQF) of the top management team in firm i in industry j in year t is computed
as follows:
X
(6) INSTRUMENT j,t = ACQ j,s,t−5 × ENFORCE INDEXs,t ,
s
where j, s, and t index industry, state, and year, respectively. ACQ j,s,t−5 is the
number of acquisitions made by established (public) companies in industry j in
state s in year t–5. The information on mergers and acquisitions required to con-
struct this variable is collected from the SDC Mergers and Acquisitions Database.
The 5-year lag allows for the expiration of retention contracts that work as “golden
handcuffs” for managers, and thus ACQ j,s,t−5 works as a measure of the supply of
managers from state s in industry j in year t.
ENFORCE INDEXs,t is the index measuring the enforceability of non-
compete agreements across different U.S. states based on Garmaise (2011),
which ranges from 0 to 1.35 Higher (lower) values of ENFORCE INDEX
indicate weaker (greater) enforceability of noncompete clauses and thus
greater (weaker) mobility of managers. The multiplication term, ACQ j,s,t−5 ×
ENFORCE INDEXs,t , therefore proxies for the supply of managers who are able
to move across firms and available for hire from state s in industry j in year t. We
then aggregate this variable at the industry-year level across states and use this as
31
In an earlier (working paper) version of their article, Ewens and Marx (2018) show that the
number of acquisitions in an industry is strongly correlated with the movement of top managers across
firms in that industry.
32
Because these noncompete clauses become operational only when top managers leave their prior
firms, the enforceability of these noncompete clauses can be thought of as a measure of the friction
facing top managers when they attempt to join the current firm.
33
Bishara, Martin, and Thomas (2015) analyze an extensive sample of CEO employment contracts
and show that 80% of these contracts contain noncompete clauses, often with a broad geographic
scope. A growing body of work (e.g., Garmaise (2011), Marx et al. (2009)) shows that higher enforce-
ability of these noncompete clauses constrains employees’ mobility (including that of managers).
34
One possible argument against our IV satisfying the exclusion restriction is that the noncompete
clauses in a state may affect the movement of inventors to the sample firm (as well as that of top
managers), thereby affecting the extent of innovation by the firm. This argument, however, is in fact
not valid because our IV is not driven by the enforceability of noncompete clauses in a state alone.
The number of acquisitions in the sample firm’s industry and state is unlikely to affect the movement
of inventors across firms. We have empirically verified that the movement of inventors across firms is
not correlated with our IV.
35
Garmaise (2011) develops an index to measure the enforceability of noncompete clauses by
considering 12 questions analyzed by Malsberger (2004), which is the central resource describing
noncompetition law in the 50 U.S. states and the District of Columbia, and assigning 1 point to each
jurisdiction for each question if the jurisdiction’s enforcement of that dimension of noncompetition
law exceeds a certain threshold, with possible totals therefore ranging from 0 to 12. Higher values
of Garmaise’s (2011) index indicate higher enforceability of the noncompete agreements in this state
and thus less mobility of the managers from this state. For example, Garmaise’s index (2011) is equal
to 0 for California and is equal to 9 for Florida after 1997. The ENFORCE INDEX used here is
constructed as the difference between 12 and the value of Garmaise’s (2011) index scaled by 12, and
thus it potentially ranges from 0 to 1.
https://doi.org/10.1017/S0022109018001497
an instrument for top management quality in a firm in industry j in year t. We ag-

gregate our instrumental variable across all 50 U.S. states to take into account the
fact that the market for top managers is likely to be a nationwide market. Further,
given that noncompete clauses have a broad geographic scope (i.e., they apply
even outside the state of the managers’ original employer), our IV is likely to be
a good proxy for the supply of top managers available to a sample firm from all
over the United States. We expect our instrument to be positively correlated with
our top management quality factor (MQF).
To instrument for the management quality (MQF) of firm i in industry j in
year t, we therefore run the following first-stage regression:
(7) MQFi,t = α + βINSTRUMENT j,t + δACQ j,t−5

+ γ Z i,t + IND FE + YR FE + ST FE + i,t .
In both the first and second stages of our IV regressions, we include the total
number of acquisitions in the sample firm’s industry 5 years prior (ACQ j,t−5 ) to
control for the effect of an industry-wide shock (e.g., merger waves) on innova-
tion. We also include fixed effects for the state in which the firm is headquartered
to alleviate the concern that the relation between top management quality and in-
novation may be driven by other state-level factors. We expect the instrument to
be positively and significantly related to our MQF, thus satisfying the relevance
condition required for a valid instrument. However, even though we control for the
direct effect of merger waves on innovation, it may be argued that the requirement
that the acquisitions in the industry in previous years be correlated with future
innovation only through the supply of managers may not hold. In this latter sce-
nario, the exclusion restriction for a valid instrument will not be satisfied. Given
this, the results of our IV analysis should be viewed as suggestive and interpreted
with caution.
Column 1 of Table 13 reports the results of the first stage of our IV analysis.
The coefficient of the instrument is positive (as predicted) and is statistically sig-
nificant at the 1% level. The first-stage F-statistic is 40.09, which is significant at
the 1% level. These findings confirm that the relevance condition for the instru-
ment is satisfied. Columns 2–4 report the second-stage results of our IV (2-stage
least squares (2SLS)) regressions using 1-, 2-, and 3-year-ahead patent counts
as our dependent variables.36 In summary, even after controlling for the poten-
tial endogeneity between MQF and innovation using our IV analysis, our MQF
still has a positive and significant impact on firms’ patent counts, total number
of citations, and citations per patent in all the specifications. We also conduct IV
analyses for innovative efficiency as well as for our innovation strategy measures
using the same instrumental variable as described here. In untabulated results, we
find that the coefficients of MQF in these 2-stage regressions are also positive and
significant.
36
We report our second-stage results using the total number of citations and the number of citations
per patent as dependent variables, respectively, in Table IA-6 of our Supplementary Material and
obtain similar results.
https://doi.org/10.1017/S0022109018001497
TABLE 13
The Effect of Management Quality on Corporate Innovation: Instrumental Variable Analysis
Table 13 reports the instrumental variable (IV) regression results of corporate innovation on the management quality factor
(MQF). The instrument is described in Section VI. Column 1 reports the first-stage regression result, regressing MQF on
the instrument and other controls. Columns 2–4 report the second-stage regression results using the number of patents
filed by a firm in a given year as the dependent variable. ln(PATENT) is the natural logarithm of 1 plus the truncation-
adjusted number of patents filed by a firm in a given year. ACQt −5 is the total number of acquisitions in the sample firm’s
industry 5 years prior. Other control variables are defined as in Table 3. Constants; year fixed effects, 2-digit SIC industry
fixed effects, and state fixed effects are included in all regressions. All standard errors are adjusted for clustering at the
firm level and are reported in parentheses below the coefficient estimates. *, **, and *** indicate statistical significance
at the 10%, 5%, and 1% levels, respectively.
MQF ln(PATENT)t +1 ln(PATENT)t +2 ln(PATENT)t +3

Variable 1 2 3 4
INSTRUMENT 0.011***
(0.002)
MQF 0.610*** 0.362*** 0.239
(0.149) (0.139) (0.161)
ACQt −5 −0.013*** 0.000 0.001** 0.002**
(0.002) (0.000) (0.000) (0.001)
ln(ASSETS) 0.069*** 0.111*** 0.125*** 0.128***
(0.008) (0.012) (0.010) (0.011)
MB 0.167*** 0.031 0.076*** 0.097***
(0.017) (0.028) (0.026) (0.031)
ROA −0.195*** 0.079** 0.045 0.021
(0.032) (0.037) (0.033) (0.035)
CAPEX/ASSETS −0.729*** 0.325** 0.155 0.089
(0.095) (0.136) (0.120) (0.124)
R&D/ASSETS 0.213*** 0.052 0.082 0.072
(0.058) (0.059) (0.051) (0.051)
RET −0.049*** −0.005 −0.013 −0.010
(0.007) (0.009) (0.009) (0.009)
AVG_TENURE −0.036*** 0.020*** 0.011** 0.006
(0.003) (0.006) (0.005) (0.006)
HHI −0.100 0.230 0.194 0.155
(0.274) (0.240) (0.220) (0.197)
No. of obs. 25,945 25,945 23,096 20,119
Adj. R 2 0.107
Industry, year, and state fixed effects Yes Yes Yes Yes
VII. Conclusion
In this paper, we use panel data on top management characteristics and a
management quality factor constructed using common factor analysis on indi-
vidual measures of management quality (management team size, the fraction of
managers with MBA degrees, the fraction of managers with Ph.D. degrees, the
fraction of members with prior work experience on the top management team,
the average number of prior board positions that each manager has held, and the
average employment- and education-based connections of each manager in the
management team) to analyze the relation between firm top management quality
and corporate innovation input and output. We show that top management qual-
ity is an important determinant of corporate innovation, with different individual
aspects of management quality affecting innovation in younger and older firms
differently. We further show that firms with higher top management quality en-
gage in more risky (“explorative”) innovation strategies. Finally, we show that
hiring more and higher-quality inventors is an important channel through which
firms with higher top management quality achieve greater innovation output.
https://doi.org/10.1017/S0022109018001497
Supplementary Material
Supplementary Material for this article is available on the authors’ Web sites:
https://www2.bc.edu/thomas-chemmanur/.
References
Acemoglu, D., and D. Autor. “Lectures in Labor Economics.” Available at http://economics.mit.edu/
files/4689 (2011).
Adams, R. B.; H. Almeida; and D. Ferreira. “Powerful CEOs and Their Impact on Corporate Perfor-
mance.” Review of Financial Studies, 18 (2005), 1403–1432.
Aghion, P.; J. Van Reenen; and L. Zingales. “Innovation and Institutional Ownership.” American Eco-
nomic Review, 103 (2013), 277–304.
Atanassov, J. “Do Hostile Takeovers Stifle Innovation? Evidence from Antitakeover Legislation and
Corporate Patenting.” Journal of Finance, 68 (2013), 1097–1131.
Balsmeier, B.; L. Fleming; and G. Manso. “Independent Boards and Innovation.” Journal of Financial
Economics, 123 (2017), 536–557.
Becker, G. S. Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Edu-
cation. Chicago, IL: University of Chicago Press (1964).
Ben-Porath, Y. “The Production of Human Capital and the Life Cycle of Earnings.” Journal of Political
Economy, 75 (1967), 352–365.
Bernstein, S. “Does Going Public Affect Innovation?” Journal of Finance, 70 (2015), 1365–1403.
Bertrand, M., and A. Schoar. “Managing with Style: The Effect of Managers on Firm Policies.” Quar-
terly Journal of Economics, 118 (2003), 1169–1208.
Bishara, N. D.; K. J. Martin; and R. S. Thomas. “An Empirical Analysis of Noncompetition Clauses
and Other Restrictive Postemployment Covenants.” Vanderbilt Law Review, 68 (2015), 1–51.
Bloom, N., and J. Van Reenen. “Measuring and Explaining Management Practices across Firms and
Countries.” Quarterly Journal of Economics, 122 (2007), 1351–1408.
Brav, A.; W. Jiang; S. Ma; and X. Tian. “How Does Hedge Fund Activism Reshape Corporate Inno-
vation?” Journal of Financial Economics, 130 (2018), 237–264.
Chemmanur, T. J., and Y. Jiao. “Dual Class IPOs: A Theoretical Analysis.” Journal of Banking and
Finance, 36 (2012), 305–319.
Chemmanur, T. J.; L. Kong; K. Krishnan; and Q. Yu. “Human Capital, Management Quality, and Firm
Performance.” Working Paper, Boston College (2017).
Chemmanur, T. J.; E. Loutskina; and X. Tian. “Corporate Venture Capital, Value Creation, and Inno-
vation.” Review of Financial Studies, 27 (2014), 2434–2473.
Chemmanur, T. J., and I. Paeglis. “Management Quality, Certification, and Initial Public Offerings.”
Journal of Financial Economics, 76 (2005), 331–368.
Chemmanur, T. J.; I. Paeglis; and K. Simonyan. “Management Quality, Financial and Investment Poli-
cies, and Asymmetric Information.” Journal of Financial and Quantitative Analysis, 44 (2009),
1045–1079.
Chemmanur, T. J.; I. Paeglis; and K. Simonyan. “Management Quality and Antitakeover Provisions.”
Journal of Law and Economics, 54 (2011), 651–692.
Chemmanur, T. J., and X. Tian. “Do Antitakeover Provisions Spur Corporate Innovation? A Re-
gression Discontinuity Analysis.” Journal of Financial and Quantitative Analysis, 53 (2018),
1163–1194.
Cohen, L.; A. Frazzini; and C. Malloy. “The Small World of Investing: Board Connections and Mutual
Fund Returns.” Journal of Political Economy, 116 (2008), 951–979.
Custodio, C.; M. A. Ferreira; and P. P. Matos. “Do General Managerial Skills Spur Innovation?” Man-
agement Science, 65 (2019), 459–476.
Eberhart, A. C.; W. F. Maxwell; and A. R. Siddique. “An Examination of Long-Term Abnormal Stock
Returns and Operating Performance Following R&D Increases.” Journal of Finance, 59 (2004),
623–650.
Ederer, F., and G. Manso. “Is Pay for Performance Detrimental to Innovation?” Management Science,
59 (2013), 1496–1513.
Engelberg, J.; P. Gao; and C. A. Parsons. “The Price of a CEO’s Rolodex.” Review of Financial Studies,
26 (2013), 79–114.
Ewens, M., and M. Marx. “Founder Replacement and Startup Performance.” Review of Financial
Studies, 31 (2018), 1532–1565.
https://doi.org/10.1017/S0022109018001497
Ewens, M., and M. Rhodes-Kropf. “Is a VC Partnership Greater Than the Sum of Its Partners?” Jour-
nal of Finance, 70 (2015), 1081–1113.
Faleye, O.; R. Hoitash; and U. Hoitash. “The Costs of Intense Board Monitoring.” Journal of Financial
Economics, 101 (2011), 160–181.
Ferreira, D.; G. Manso; and A. C. Silva. “Incentives to Innovate and the Decision to Go Public or
Private.” Review of Financial Studies, 27 (2014), 256–300.
Finkelstein, S., and D. C. Hambrick. Top Executives and Their Effects on Organizations. Minneapolis/
St. Paul, MN: West (1996).
Gardner, H. “Frame of Mind: The Theory of Multiple Intelligences.” New York, NY: Basic Books
(1983).
Garmaise, M. J. “Ties That Truly Bind: Noncompetition Agreements, Executive Compensation, and
Firm Investment.” Journal of Law, Economics, and Organization, 27 (2011), 376–425.
Griliches, Z. “Patent Statistics as Economic Indicators: A Survey.” Journal of Economic Literature,
28 (1990), 1661–1707.
Griliches, Z.; A. Pakes; and B. H. Hall. “The Value of Patents as Indicators of Inventive Activity.”
NBER Working Paper No. 2083 (1988).
Gu, F. “Innovation, Future Earnings, and Market Efficiency.” Journal of Accounting, Auditing and
Finance, 20 (2005), 385–418.
Hall, B. H.; A. B. Jaffe; and M. Trajtenberg. “The NBER Patent Citation Data File: Lessons, Insights
and Methodological Tools.” NBER Working Paper No. 8498 (2001).
Hirshleifer, D.; P.-H. Hsu; and D. Li. “Innovative Efficiency and Stock Returns.” Journal of Financial
Economics, 107 (2013), 632–654.
Hirshleifer, D.; A. Low; and S. H. Teoh. “Are Overconfident CEOs Better Innovators?” Journal of
Finance, 67 (2012), 1457–1498.
Hirshleifer, D., and A. V. Thakor. “Managerial Conservatism, Project Choice, and Debt.” Review of
Financial Studies, 5 (1992), 437–470.
Kaplan, S. N.; M. M. Klebanov; and M. Sorensen. “Which CEO Characteristics and Abilities Matter?”
Journal of Finance, 67 (2012), 973–1007.
Kogan, L.; D. Papanikolaou; A. Seru; and N. Stoffman. “Technological Innovation, Resource Alloca-
tion, and Growth.” Quarterly Journal of Economics, 132 (2017), 665–712.
Lanjouw, J. O., and M. Schankerman. “Patent Quality and Research Productivity: Measuring Innova-
tion with Multiple Indicators.” Economic Journal, 114 (2004), 441–465.
Lerner, J. “The Importance of Patent Scope: An Empirical Analysis.” RAND Journal of Economics,
25 (1994), 319–333.
Lerner, J.; M. Sorensen; and P. Strömberg. “Private Equity and Long-Run Investment: The Case of
Innovation.” Journal of Finance, 66 (2011), 445–477.
Lerner, J., and J. Wulf. “Innovation and Incentives: Evidence from Corporate R&D.” Review of Eco-
nomics and Statistics, 89 (2007), 634–644.
Li, G.-C.; R. Lai; A. D’Amour; D. M. Doolin; Y. Sun; V. I. Torvik; A. Z. Yu; and L. Fleming. “Dis-
ambiguation and Co-Authorship Networks of the U.S. Patent Inventor Database (1975–2010).”
Research Policy, 43 (2014), 941–955.
Malmendier, U., and G. Tate. “CEO Overconfidence and Corporate Investment.” Journal of Finance,
60 (2005), 2661–2700.
Malsberger, B. M. “Covenants Not to Compete: A State-by-State Survey.” Washington, DC: BNA
(2004).
Manso, G. “Motivating Innovation.” Journal of Finance, 66 (2011), 1823–1860.
Marx, M.; D. Strumsky; and L. Fleming. “Mobility, Skills, and the Michigan Non-Compete Experi-
ment.” Management Science, 55 (2009), 875–889.
Rajan, R. G., and L. Zingales. “Financial Dependence and Growth.” American Economic Review, 88
(1998), 559–586.
Sah, R. K., and J. E. Stiglitz. “The Architecture of Economic Systems: Hierarchies and Polyarchies.”
American Economic Review, 76 (1986), 716–727.
Sah, R. K., and J. E. Stiglitz. “The Quality of Managers in Centralized versus Decentralized Organi-
zations.” Quarterly Journal of Economics, 106 (1991), 289–295.
Sapra, H.; A. Subramanian; and K. V. Subramanian. “Corporate Governance and Innovation: Theory
and Evidence.” Journal of Financial and Quantitative Analysis, 49 (2014), 957–1003.
Seru, A. “Firm Boundaries Matter: Evidence from Conglomerates and R&D Activity.” Journal of
Financial Economics, 111 (2014), 381–405.
Sørensen, J. B., and T. E. Stuart. “Aging, Obsolescence, and Organizational Innovation.” Administra-
tive Science Quarterly, 45 (2000), 81–112.
https://doi.org/10.1017/S0022109018001497
Spiegel, M. I., and H. Tookes. “Identifying an IPO’s Impact on Rival Firms.” Working Paper, Yale
University (2016).
Stein, J. C. “Takeover Threats and Managerial Myopia.” Journal of Political Economy, 96 (1988),
61–80.
Stein, J. C. “Efficient Capital Markets, Inefficient Firms: A Model of Myopic Corporate Behavior.”
Quarterly Journal of Economics, 104 (1989), 655–669.
Tian, X., and T. Y. Wang. “Tolerance for Failure and Corporate Innovation.” Review of Financial
Studies, 27 (2014), 211–255.

Chemmanur Et Al 2019

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chemmanur Et Al 2019

Uploaded by

Copyright:

Available Formats

https://doi.org/10.

Top Management Human Capital, Inventor

*Chemmanur (corresponding author), chemmanu@bc.edu, Boston College Carroll School of

quality) involves subjective notions of what constitutes a higher-quality manage-

or in more conservative “exploitative” innovative strategies (where they may pur-

II. Theory and Hypothesis Development

III. Data and Sample Selection

2. Common Factor Analysis on Individual Measures of Management Quality

individual measures of management quality using as few factors as possible. Next,

(1) MEASUREi,t = α[ln(FIRM SIZE)i,t ] + β[ln(FIRM SIZE)i,t ]2

FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5 FACTOR 6 FACTOR 7

Variable Loadings on First Factor Correlation with First Factor Communality

Panel C. Summary Statistics of the First Factor

for a management team, where total compensation, in addition to cash com-

ln(AVG_TOT_COMP) ln(AVG_CASH_COMP) EQUITY/TOT_COMP

expectation, the coefficients on MQF are positive and statistically significant at

of non-self-citations (i.e., the number of times that a firm’s patent portfolio in

IV. Empirical Tests and Results

(3) R&D/ASSETSi,t+n = α + βMQFi,t + γ Zi,t + IND FE + YR FE + i,t ,

where i indexes firm; t indexes time; and n equals 1, 2, or 3. Z is a vector of control

R&D/ASSETSt +1 R&D/ASSETSt +2 R&D/ASSETSt +3

MQF 0.008*** 0.008*** 0.007***

B. The Effect of Management Quality on Corporate Innovation

(4) INNOV OUTPUTi,t+n = α + βMQFi,t + γ Z i,t

where the dependent variables (INNOV OUTPUT) include ln(PATENT),

Panel A. The Effect of MQF on the Number of Patents

MQF 0.016*** 0.016*** 0.016***

Panel C. The Effect of MQF on the Number of Citations per Patent

ln(CPP)t +1 ln(CPP)t +2 ln(CPP)t +3

MQF 0.067** 0.084***

Chemmanur, Kong, Krishnan, and Yu

Difference Difference Difference Difference Difference Difference Difference Difference

Panel A. The Effect of MQF on Explorative and Exploitative Patents

EXPLOREt +1 EXPLOITt +1 EXPLOREt +2 EXPLOITt +2 EXPLOREt +3 EXPLOITt +3

Difference Difference Difference

and moderately successful innovations (M CITE), based on their citation distri-

Difference Difference Difference

MQF 0.303*** 0.147*** 0.144*** 0.156*** 0.159*** 0.003

Panel B. The Effect of MQF on Non-Self-Citations and Self-Citations

NS_CITEt +1 S_CITEt +1 NS_CITEt +2 S_CITEt +2 NS_CITEt +3 S_CITEt +3

MQF 0.316*** 0.253*** 0.313*** 0.251*** 0.304*** 0.244***

Difference 0.062*** 0.062*** 0.060***

DIVERSITYt +1 DIVERSITYt +2 DIVERSITYt +3 3YR_DIVERSITYt +1

MQF 0.036*** 0.039*** 0.036*** 0.041***

likely to engage in explorative innovation strategies, those involving new tech-

V. Analysis of the Inventor Mobility Channel

(5) NET INFLOWi,t+n = α + βMQFi,t + γ Zi,t + IND FE

where i indexes firm; t indexes time; and n equals 1, 2, or 3. Z is a vector of

NET_INFLOWt +1 NET_INFLOWt +2 NET_INFLOWt +3

NET_INFLOW_ NET_INFLOW_ NET_INFLOW_ NET_INFLOW_ NET_INFLOW_ NET_INFLOW_

Difference Difference Difference

Difference 0.050*** 0.049*** 0.052***

significance of the difference between the coefficients on MQF for high-quality

ln(PATENT)t +1 ln(PATENT)t +2 ln(PATENT)t +3

ln(CITE)t +1 ln(CITE)t +2 ln(CITE)t +3

ln(CPP)t +1 ln(CPP)t +2 ln(CPP)t +3

VI. Instrumental Variable Analysis

an instrument for top management quality in a firm in industry j in year t. We ag-

(7) MQFi,t = α + βINSTRUMENT j,t + δACQ j,t−5

MQF ln(PATENT)t +1 ln(PATENT)t +2 ln(PATENT)t +3

You might also like

(3) R&D/ASSETSi,t+n = α + βMQFi,t + γ Zi,t + IND FE + YR FE + i,t ,

MQF 0.008* 0.008* 0.007***

MQF 0.016* 0.016* 0.016***

MQF 0.067 0.084*

MQF 0.303* 0.147* 0.144* 0.156* 0.159*** 0.003

MQF 0.316* 0.253* 0.313* 0.251* 0.304* 0.244*

Difference 0.062* 0.062* 0.060***

MQF 0.036* 0.039* 0.036* 0.041*

Difference 0.050* 0.049* 0.052***