Content uploaded by Heike Grimm
Author content
All content in this area was uploaded by Heike Grimm on Apr 23, 2015
Content may be subject to copyright.
Testing the causal relationship between academic
patenting and scientific publishing in Germany:
Crowding-out or reinforcement?
Heike M. Grimm •Johannes Jaenicke
Springer Science+Business Media New York 2014
Abstract The paper investigates the intertemporal spillover effects from patenting to
future publishing activities and vice versa among university employees with a country
focus on the German Laender Bavaria, Saxony and Thuringia. Individual data from uni-
versity patentees who successfully issued a patent at a public university before and after
2005 from the selected German Laender is used for measuring the Granger-causal effects
between both activities. The interaction of personal and institutional characteristics of
academic patentees is taken into account. By using Granger-causality tests in a dynamic
panel model, we test the overall effect as well as group or Laender specific effects. Our
findings show that there is a positive feedback relationship between patenting and pub-
lishing activities. An increase in patent applications results in higher numbers of future
publications; reciprocally, an increase of publications contributes to a higher output of
future patent applications. Additionally, we find interrelations of the research output with
seniority, academic degree of the scientists and non-university work experience. The paper
further presents findings about motives, skills and experience of so-called star scientists
and other academic inventors.
Keywords Academic patenting Scientific publishing Academic entrepreneurship
Academic policy Star scientists Germany Granger-causality
JEL Classification O31 I23 C33
H. M. Grimm (&)
Willy Brandt School of Public Policy, University of Erfurt, Nordha
¨user Straße 74, Building 39,
99089 Erfurt, Germany
e-mail: heike.grimm@uni-erfurt.de
J. Jaenicke
Hamburg Institute of International Economics (HWWI), Regional Office Thuringia, Elisabethstraße
10, 99096 Erfurt, Germany
e-mail: jaenicke@hwwi.org
URL: http://www.hwwi.org
123
J Technol Transf
DOI 10.1007/s10961-014-9353-z
1 Introduction
1.1 The policy debate
With the alteration of the German Employee Invention Act, in 2002, the German Ministry
of Education and Research acknowledged the creation, exploitation and diffusion of uni-
versity innovation as an important determinant for economic development (BMBF 2001,
2002). As a result, it implemented new forms of academic management and governance to
facilitate the patenting process and to promote the entrepreneurial commercialization of
academic inventions at universities which gained in importance not only as drivers of
technological advancement but also as crucial steps to enter the entrepreneurial route
(Landry et al. 2007,2010; Audretsch and Aldridge 2009; D’Este et al. 2012; Audretsch
2013). In other words, it approved entrepreneurship not only as a route of discovery but
also as the capability to exploit an academic innovation in a profitable manner (Shane
2002; Wright et al. 2004; Mowery and Sampat 2005; Rothaermel et al. 2007; Bishop et al.
2011).
The new public policy has not only been welcomed but also criticized by German
academia because it emphasizes the role and importance of increasing commercialization
of innovations generated at public universities. Criticism refers to the possible crowding-
out effect and decrease of long-term and basic research; a crowding-out effect concerning
publications for the benefit of increased patents and associated financial benefits; a
growing, direct influence of industry and government on research agendas at universities;
increasing pressure on academics with traditional research habits and a focus on basic
research to carry out more applied-oriented research in the future (Geuna 2001; Geuna and
Nesta 2006; Klitkou and Gulbrandsen 2009). These arguments contribute to a rather
skeptical view with respect to the emergence of new public policy to increase patenting
and the potential commercialization of a patent.
1.2 Objectives of this study
The above-mentioned German policy initiative aimed at offering better organizational,
financial and legal conditions for patenting activity at universities paving the way for
academic entrepreneurship. Regardless of the organizational and legal environment, we
argue that the individual determinants of scientists to patent matter considerably as well as
their disciplinary and institutional background. The German government presumed that the
motivation to patent or profit from patents by using it commercially is strongly related to
better legal and organizational conditions disregarding that ‘‘we know remarkably little
about who in the university is patenting’’ (Stephan et al. 2007, p. 72). Various authors have
recently pinpointed to the academic void that our knowledge about the personal charac-
teristics and context of academic patentees is insufficient (Walter et al. 2013; Wright 2013;
Hu
¨lsbeck et al. 2013; Welter 2011). This is specifically true for the German context. What
are the personal characteristics of academics who patent and what is their institutional
environment? Further, the crucial question whether the new policy approach may result in
a trade-off between patenting and publishing has not been tackled in Germany empirically
resulting in a controversial debate about the consequence of policy to better promote
academic patenting. The following paper, therefore, investigates the relationship between
patenting and publication activity (and vice versa) as well as the personal and institutional
conditions for academic researchers to align patenting and publishing activities. Only very
few studies have looked at the set of skills and capabilities of an academic entrepreneur
H. M. Grimm, J. Jaenicke
123
and/or patentee influencing publishing, patenting and entrepreneurial behavior with focus
on Germany (Walter et al. 2013; Grimm and Jaenicke 2012; Buenstorf 2009; Czarnitzky
et al. 2009, for other countries see Hoye and Pries 2009; Bercovitz and Feldman 2008;
Grandi and Grimaldi 2005). Czarnitzky et al. (2009, p. 33) underline that with regard to
‘‘ ( …) personal characteristics and their faculty environment (…) no systematic database
exists’’. The German entrepreneurship literature lacks profound knowledge about the skills
characterizing academics who patent, transfer new knowledge to the private sector and/or
consider entrepreneurial activity as a professional option next to an academic career.
We will therefore present a comprehensive database which allows us to profile German
academic patentees in a first step. We further present a Sect. which differentiates between
so-called star scientists (Zucker and Darby 1996), who have an excellent record in pat-
enting or publishing or in both disciplines, and their colleagues because excellence in
patenting as well as publishing turned out to play an important role for complementing
both activities. We also assume that this cohort of academic star scientists has a vested
interest in knowledge transfer activities. As we will illustrate in Sect. 2, there is increasing
evidence that scientific excellence is expressed by the reinforcement of patenting and
publishing. This relationship is analyzed with Granger-causality tests in heterogeneous
panels and discussed in this paper.
Our major research goals are twofold: (1) to examine the relationship between patenting
and publishing and vice versa; (2) to examine whether individual and institutional char-
acteristics influence scientist’s publication and patenting activity. Further, we examine so-
called star scientists’ profile in the context of publication and patenting activity as well as
entrepreneurial behavior.
Based on the information and findings of a survey amongst university patentees who
successfully applied a patent in 2005 in three German Laender (Thuringia, Saxony and
Bavaria), we will investigate the influence of selected variables on patenting and pub-
lishing behavior of academic patentees. We have selected the variables ‘‘gender’’, ‘‘age’’,
‘‘academic degree’’ (no PhD, PhD, Habilitation which also serves as a proxy for seniority),
‘‘non-university working experience’’ (which serves as a proxy for a vested interest in
collaboration with the private sectors), ‘‘university type’’ (university, university of applied
sciences) and ‘‘discipline’’ (construction, engineering, electrical engineering, biology,
chemistry, physics, computer science, medicine) and empirically assessed their influence
on patenting and publishing activity. We have further explored the influence of interacting
independent control variables on the lagged dependent variables patenting and publishing.
With reference to the questionnaire design, the year 2005 serves as a time threshold for
patent applications and publications.
We are able to demonstrate with our findings a one-directional Granger-causal link
running from patenting to publishing and vice versa. This implies that there exists a
feedback relationship between both research activities. In the case of Saxony, more pub-
lications will increase the number of future patent applications. More patent applications
will increase the number of future publications of all patentees up to the age of 56 across
all Laender. Thus, on average, we are not able to find a crowding-out effect. Personal
characteristics (such as age), disciplines and academic degrees (such as Habilitation)
additionally play an important role explaining patenting and publishing activities.
The paper is structured as follows. After having presented the theoretical background in
Sect. 2, we explain the method, data selection and handling in Sect. 3. Sect. 4summarizes
descriptive findings of our empirical research and Sect. 5the results of the Granger-
causality test in a dynamic panel model. In Sect. 6we present a profile of selected star
Crowding-out or reinforcement?
123
scientists in the context of their publication and patenting activity. Sect. 7summarizes our
findings, and presents recommendations for academic policy.
2 Theoretical framework: the patenting–publishing relationship re-considered
The research output or research productivity of university academics is measured by the
number and quality of publications, on the one hand, and patenting activities including
their application and/or commercialization, on the other hand (Baldini et al. 2007; Lowe
and Gonzalez-Brambilla 2005). Evaluations about the relationship between patenting and
publication activities of university researchers point to contradictory findings whereas
complementary effects are more often cited than crowding-out effects. Academic patenting
can implicate that a (patent-related) publication will be postponed or not pursued which
turns out to be a trade-off for many academics whose reputation and research output is
measured by the number and quality of publications across many disciplines. Although
most studies present evidence that patenting and publishing are related activities and ‘‘that
patent and publication productivity are positively correlated at both, the individual and
institutional levels, the nature of this relationship and its implications for research agendas
remains elusive’’ (Moutinho et al. 2007, p. 358).
2.1 Patenting-publishing duality
Therefore, it comes as a surprise that a remarkable number of researchers assessed a rein-
forcement effect between academic patenting and scientific publishing in the US (Powers and
McDougall 2005; Azoulay et al. 2007,2009; Calderini et al. 2007; Stephan et al. 2007,
Fabrizio and Di Minin 2008; Geuna and Mowery 2007). Azoulay et al. (2009) concluded that
patenting has a positive effect on the rate of publications and a weak positive effect on the
quality of publications after examining a panel dataset of 3,862 academic life scientists.
Academic patentees publish more than academic non-patentees up by over a third. Earlier
research by Stephan et al. (2007) examined the cross-sectional nature and relationship
between patenting and publishing while exploiting the 1995 Survey of Doctorate Recipients
including 10,962 doctoral scientists and engineers. Their results also confirm that the two
research outputs are strongly complementary. Among other results they highlight that
environmental framework conditions matter, among them, the characteristics of the insti-
tution employing the researcher and its history of patenting activity, as well as the specific
field of research. Both criteria positively influence the researchers’ attitude to patent. Fabrizio
and Di Minin (2008) presented similar conclusions when matching a sample of 166 academic
patentees with non-patentees. In some specifications, a statistically significant and positive
numerical effect of patenting on publishing was assessed. In an early study, Murray and Stern
(2007) arrived at the conclusion that most researchers in the field of biotechnology patent and
publish their research and findings in a consistent manner. Meyer (2006) presented a bib-
liometric analysis about patenting behaviors of scientists in the fields of nano-science and
nano-technology, and concluded that academic patentees outperform their non-patenting
colleagues in terms of quantity and quality of publications. In a European context, the
following findings point to a duality of patenting and publishing. Breschi et al. (2008)
assessed in an empirical study that Italian academics who apply successfully a patent with the
European Patent Office publish more than the control group (non-applying professors).
Crespi et al. (2011) assessed a publishing-patenting complementarity for UK scientists from
two major London universities up to a certain level of patenting output. Carayol (2007) found
H. M. Grimm, J. Jaenicke
123
that patents are positively related to publications in France. For Belgium, van Looy et al.
(2006) present findings which re-confirm a positive patenting-publishing relationship. The
comparison between academic patentees and non-patentees revealed that the first group
published up by over a third; as mentioned above, a similar finding was presented by Azoulay
et al. (2009) for US researchers. Czarnitzky et al. (2009) represented an empirical study about
the patenting and publishing behaviors of 1,876 German patenting professors with more than
34,000 publications, holding about 7,000 patents. They distinguished between a professorial
cohort whose patents were assigned to corporations, and a cohort whose patents were
assigned to non-profit organizations (such as universities). They conclude that patents in
collaboration with non-profit organizations contribute to the quality and quantity of scientific
publications; patent co-operations with private sector organizations rather result in publi-
cation activities characterized of lower quality and to some extent also lower quantity. These
findings re-confirm former evidence presented by Fabrizio and Di Minin (2008) who also
assessed that university patents complement publications if patents are assigned to
universities.
In a nutshell, there is striking empirical evidence that especially top researchers both
publish and patent, and that the publication output is not negatively affected due to fre-
quent patent activities (Crespi et al. 2011). It is, therefore, specifically interesting to assess
the patenting and publishing activities of top researchers and to investigate their personal
characteristics. Scientists who well-balance publishing and patenting are rather categorized
as ‘‘stars’’ who outperform their non-patenting colleagues with regard to outcome and
quality of publications (Czarnitzky et al. 2009, p. 33).
2.2 Patenting–publishing trade-offs
The substitution effect (patenting suppressing publishing in the long run) is observed less
frequently. The studies by Blumenthal et al. (1996), Campbell et al. (2002) and Krimsky
(2003) presented evidence on delayed publications caused by patent activities and less
often crowding-out effects towards commercial interests. Klitkou and Gulbrandsen (2009,
p. 96) assessed that patenting activity implied a delay of publishing or the placement of
publications in less prestigious journals. Agrawal and Henderson (2002) published an often
cited study with mixed findings. The researchers selected 236 scientists in two departments
of the Massachusetts Institute of Technology (MIT) applying estimated fixed-effect
regressions of the effect of patenting within a 15-year time frame. They concluded that
patenting had neither a substituting nor complementary influence on publishing rates
though ‘‘most patentable research is also publishable’’ (ibd., p. 58). However, they came to
the conclusion that increased patenting activity correlated with increased rates of citations
of publications placed by faculty members.
2.3 Possible factors influencing patenting and publishing activities
According to relevant and above cited literature, the possible individual and institutional
characteristics and skills that have an influence on academic patenting and publishing as
well as entrepreneurial behavior are age or seniority, gender, discipline, academic degree,
academic position, university type, collaboration with industry and an academic career
outside of university (non-university working experience) (Gulbrandsen 2005; Audretsch
and Aldridge 2009; Shin and Cummings 2010; D’Este et al. 2012). Klitkou and Gul-
brandsen (2009, pp. 94/95) distinguished between three important contextual aspects which
need to be considered to better understand why some academics stick to the either-or
Crowding-out or reinforcement?
123
option while others consider patenting as part of their research agenda next to publications:
(1) the national and regulatory framework matters; (2) the institutional environmental
conditions which are relevant including an entrepreneurial culture, the appreciation of
excellent performance or the support through technology transfer office (TTOs); (3) the
academic discipline and technological specialization which have to be taken into account
to benchmark the importance and role of patenting and commercialization. The pre-con-
ditions, motivations and opportunities to align academic publishing with patenting and
commercialization, are significantly correlated with various context variables, among them
personal and contextual variables. With regard to personal aspects, Carayol (2007)
assessed in a study at a French University that seniority had an influence on patenting
activity. While younger researchers concentrated on the publication output, elder peers
showed an increasing interest in patenting resulting in fewer publications (also confirmed
by Stephan et al. 2007). Moutinho et al. (2007) also confirmed that age, job stability and an
advanced career are associated with previous patenting. Their findings draw on a survey
with Portuguese scientists in life sciences and biotechnology working for public research
organizations. They also point to the fact that most researchers from this sample showed a
low interest in patenting based on the perception that this procedure implied a complex
process resulting in weak personal and financial benefits. The institutional conditions were
assessed as little helpful discouraging specifically young researchers to consider patenting
as research output. Kim (2008) underlined that an increased interest in patenting is pri-
marily dependent on intrinsic motives of academic researchers rather than financial
motives and rewards associated with patent application and commercialization. Baldini
et al. (2007) concluded that the decision ‘‘to patent or not to patent’’ is predominantly
motivated by personal considerations and attributes whereas contextual conditions can play
a role and can significantly be influenced by policymaking.
Agrawal and Henderson (2002) pointed to the fact that a close interaction with industry
turned out to motivate scientists and engineers towards patenting. Industry-oriented
activities and industry funding have also been identified as determinants of academic
patenting by other researcher, among them Balconi et al. (2004) with focus on Italy and
Meyer-Krahmer and Schmoch (1998) who analyzed university-industry interaction in
Germany. Moutinho et al. (2007) confirmed that research collaboration with industry and
previous working experience in industrial sector organizations positively influenced the
scientist
´s willingness and interest to patent and commercialize. Crespi et al. (2011) also
found a positive correlation between patent activity and engagement in knowledge transfer
channels with private sector organization for UK scientists. They further presented weak
evidence in terms of scientific field effects. Most basic sciences, namely physics and
chemistry, tended to cause a crowding-out of publishing in the course of increased pat-
enting activity. On the contrary, computer science and engineering complement patenting
and publishing by numbers. With regard to disciplines, Crespi et al. (2011) and Calderini
et al. (2007) underlined that patenting has different effects depending on the scientific field.
A difficulty to align patenting and publishing was observed for physicians by Stephan et al.
(2007) but also complementary effects in the field of engineering and life sciences (to a
lesser extent). In the context of publishing, Moutinho et al. (2007) underline that disci-
plinary differences are the main source of variance in faculty publication output. Hard
disciplines, above all medical and health sciences, have a very high publication output.
The research presented above pinpoints to highly interesting findings regarding personal
and institutional factors influencing the patenting–publishing relationship. Nonetheless,
this field of research is still rather piecemeal. Diverse countries have been focus of
empirical and qualitative investigations which makes it difficult to draw any general
H. M. Grimm, J. Jaenicke
123
conclusions. Therefore, we aim at contributing to this field of research by analyzing the
relationship between patenting and publishing and vice versa by specifically taking into
account the personal and, to a lower extent, institutional context variables of academic
patentees. This is the main contribution of this paper.
3 Research question, method and data
Are there any spillover effects from successful patenting activities to successful publishing
activity or vice versa? Are patenting and publishing activity positively related? To answer
the first research questions, we apply the concept of Granger-causality (Granger 1969)toa
panel of N university patentees observed over T periods using the approach of Hurlin
(2005) to account for different possibilities of heterogeneity. For each patentee ithe
number of patent applications at time t, t =1,2, x
it
, is causing the number of publications
y
i
, if we are better able to predict y
i
using all available information X, including x
i
in
comparison to prediction of y
i
using the information set Xwithout x
i
. Specifically (and
neglecting the subscript i), the prediction performance of yis measured by the minimum
mean square prediction error MSE (y|X).
(1) If MSE(y|X)\MSE(y|X\x)xis called Granger-causal to y.
(2) If MSE(y|X)=MSE(y|X\x)xis called Granger-noncausal to y.
(3) If MSE(x|X)\MSE(x|X\y)yis called Granger-causal to x.
(4) If MSE(x|X)=MSE(x|X\y)yis called Granger-noncausal to x.
(5) If xis Granger-causal to y and y is Granger-causal to x feedback occurs between
yand x
In the information set Xindividual specific, Laender and group-specific cross-sectional
information are included.
To test the hypothesis (1) versus (2), we estimate a dynamic panel
yi;t¼a0þaldlþbl;gxi;t1þcl;gyi;t1þdgziþei;t;ð1Þ
with
•Laender specific effects d
l
with l=Bavaria, Saxony, Thuringia,
•the lagged variables x
t-1
and y
t-1
number of patent applications until 2005 and number
of publications until 2005,
•the variables zcontrolling for group gspecific effects like university type, discipline,
non-university work experience, academic degree, age, gender, and other personal
characteristics,
•the innovation e,possibly correlated across Laender l,
•two time periods t=until 2005, 2006–2012,
•the patentee i, 1, 2,…,N.
The parameter b
l,g
and c
l,g
may be Laender or group-specific, or equal over all indi-
viduals. This implies that we are able to distinguish between Laender-specific, group-
specific and homogenous Granger-causal relations. In the most simple case, the Granger-
noncausality hypothesis is H
0
:b=0 against the alternative of Granger-causality H
1
:
b=0 for all individuals i =1, 2,…, N. In the case of group specific Granger-causality,
the group-specific Granger noncausality hypothesis is H
0
:b
i
=0 (for all individuals ithat
are member of the group g=1, 2,…,G) against the alternative of Granger-causality H
1
:
b
i
=0 and b
i
=b
j
is equal for group members (with i, j [gand i=j) and b
k
is zero or
Crowding-out or reinforcement?
123
not for members of other groups (k62 g). In the case of Laender-specific Granger-causality
the hypothesis are accordingly, but ldefines the group membership.
We allow for Laender-specific fixed effects in some specifications to account for het-
erogeneity within the data set. We employ seemingly unrelated panel estimation tech-
niques to account for neighborhood effects causing a cross-sectional correlation structure
of the residuals. Neglecting heterogeneity may lead to biased Granger-causality test results.
Nevertheless, for small T and small N, the test statistics may be non-standard and will
result in size distortions by applying standard Wald tests, as pointed out by Dumitrescu and
Hurlin (2012). In our case, T =2 is small but represents the entire research history of the
N=92 academic employees. The total number of observations is N 9T, but due to the
lagged x
t-1
and y
t-1
variables in the model we will end up with 92 observations. We
assume that in our case the size distortions of the Wald test are not too severe.
The reverse Granger-causal hypothesis, (3) versus (4) is tested in a new estimation
model altering x and y in the dynamic panel equation. The second research question,
crowding-out or reinforcement is analyzed by looking at the sign of the parameter.
The data source is based on several matched databases: Information from the German
Patent and Trade Mark Office (PTO), a self-administrated survey and the databases Sci-
ence Citation Index Expanded (1899-present), Social Sciences Citation Index (1956-
present) and the Arts & Humanities Citation Index (1975-present)—in the following
abbreviated by the acronym SSCI-E—provided by Thomson Reuters Web of Science.
The patent exploitation office in Thuringia provided data on patent applications by
public research institutes and universities in Bavaria, Thuringia and Saxony issued in 2005
or earlier. The information included, among others, the names of patentees, university
affiliations, and description of patent and personal addresses of patentees. Additional data
was added by using selected data published in the German Patentatlas and the German
Patent Office (Greif and Schmiedl 2006).
The herewith presented survey and study draws on all 2005 patent registrations of
public universities of these three Laender which were included in a sample survey to
profile the patentees with regard to their personal characteristics, publication and patent
activities. The survey further contributed to an assessment of new public policy in a clearly
defined space (Grimm and Jaenicke 2012).
In compliance with the 2005 data compiled by the Patent Exploitation Agencies (PEA)
Thuringia, 703 patent applications were successfully registered in Thuringia in 2005;
6.2 % of them by patentees of higher educational institutions (HEIs) in Thuringia. In 2005,
847 patents were registered in Saxony; 10.5 % of them by academic entrepreneurs from
HEIs (DPMA 2006, p. 13). In Bavaria, 13,449 patent applications were successfully
registered, from which only 0.34 % by university academics (46 in total). The number of
academic patent applications is surprisingly low and can partly be explained by the fact
that the majority of applications were accomplished by teams. All patentees (252) and
patenting teams of HEIs were included in the survey.
A questionnaire was sent out to the academic patentees, in 2009 and 2010, to ask them
for confirmation of their patenting activity but also about their motivations, incentives,
career paths, support structure at university and within the state, commercialization results,
potential collaboration with industry, and selected personal characteristics. The survey and
interview data was incorporated into the database.
1
1
In the context of the herewith presented paper and research, we took the opportunity to add missing
information by drawing on patentee’s personal or university websites to specify age and highest university
degree of the respondents which we did not receive through the questionnaires.
H. M. Grimm, J. Jaenicke
123
We received 92 responses. The overall response rate is 36.5 %, varying from 27.2 % in
Bavaria to 46.6 % in Saxony. Detailed information is provided in Table 1.
To investigate the intertemporal relations between publishing and patenting we searched
for patent applications of our respondents between 2006 and 2012 using the online access
of the PTO data base. To identify the individual respondent in this data base we use the
name and address of the respondent. In some cases, it was difficult to identify the
respondent. If the respondent meanwhile moved to another city, this strategy will result in
an underestimated number of patent applications.
To estimate the publishing activities, we used the SSCI-E index provided by Thomson
Reuters Web of Science.
2
We examined the publication records of patentees, measured by
the number of papers of a patentee included in the Web of Science until 2005, based on a
query which included the name of the patentee/author, the publication year (C2005) and
the name of university. Furthermore, we examined the future publishing success after the
patenting activity in 2005, using the publication record(s) from 2006 until the most recent
available data, in May 2012. By counting the publication records after successfully
applying for a patent in 2005, we aimed at investigating whether a patent application and/
or commercialization leads to academic distraction measured as decreasing publication
outputs. The patent activity of academic researchers is the self-reported number of patents
and does not necessarily relate to the employer university but rather to any patent activity
within and beyond the university environment (total number of patents per researcher).
This strategy limits the likelihood of confusions and wrong counting of publications. The
selection of three German Laender for in-depth analyses is based on the following con-
siderations: Due to the characteristics of the German federal system, the sixteen states
differ considerably when it comes to academic and other forms of policy-making, delivery
and outcome. Therefore, the two most innovative (measured by number of patent appli-
cations per year) and entrepreneurial (measured by entrepreneurial activity per year) New
German Laender, Thuringia and Saxony (Greif and Schmiedl 2006), were selected as cases
for empirical research.
3
In 2010 and 2011, Saxony reached a seventh-place ranking (in
2005, ranked number nine) in the German patenting statistics, Thuringia reached a tenth-
place ranking (same as in 2005). Further, Bavaria was selected as a case ranking as the
second most innovative West German state, since 2008, and as the most innovative state,
from 1996 until 2007, in the context of patent applications. From a German perspective,
the research approach is twofold: By comparing the patenting and publishing activity of
patentees from East German universities in Thuringia and Saxony, a most similar research
design is pursued. Both states are characterized by a comparable and similar socio-cultural,
political and economic historical and developmental experience. Both went successfully
through a transformational process after German unification with regard to economic but
also educational development. By adding Bavaria to the comparative analysis, the design
takes on features of a most different research design. This approach follows the hypothesis
that Bavaria—being one of the most entrepreneurial and innovative West German states—
serves as a best practice example when it comes to the promotion of innovative academics
due to highly supportive environmental university conditions and professional academic
policymaking.
2
According to the Web of Science factsheet, the SSCI-E covers over 12,000 of the highest impact journals
worldwide from 256 categories, including Open Access journals and 148,000 conference proceedings from
the most significant conferences, symposia, seminars, colloquia, workshops, and conventions worldwide.
3
The selected Laender are thoroughly profiled in Grimm and Jaenicke (2012, p. 459).
Crowding-out or reinforcement?
123
4 Descriptive statistics
The descriptive statistics of the respondents are presented in Table 2. Some patentees have
already applied for a patent in their early years of their career; others apply rather towards
the end of their career. The median age is 48 in Thuringia, 39 in Saxony, and 42 in Bavaria.
Only a very small proportion of patentees are female, over the complete sample only
6.5 %. A considerable number of respondents gained professional experience outside the
university system. The median patentee applied twice for a patent until 2005, the upper
quarter of Thuringian patentees applied 8.5 times for a patent; the upper quarter of
Bavarian and Saxonian patentees around six times. After 2005 the median patentee applied
once across the three Laender, the upper quarter patentee applied between twice (in
Bavaria) and three times (in Saxony). Additional tests for the sample until 2005 (after
2005) show that the mean number of patent applications is higher in Thuringia (Saxony) in
comparison with Bavaria and Saxony (Thuringia), but these differences are not significant.
Using the Satterthwaite (1946) and Welch (1951) test, we are able to find highly significant
Table 1 Distribution and total number of responses
Universities Population surveyed Survey respondents Response rate (%)
Thuringia 69 31 44.9
Saxony 58 27 46.6
Bavaria 125 34 27.2
Total 252 92 36.5
Table 2 Key characteristics of respondents
Patentees characteristics Thuringia Saxony Bavaria
Age (1st/2nd/3rd quartile) 32.5/48/62 31.5/39/48 36/42/54
Gender (percentage female) 6.5 7.4 5.9
Highest university degree
B.A., M.A. or equivalent 11 7 1
PhD 10 11 17
Post lecturer qualification (Habilitation) 10 9 16
Prior employment
Private sector 11 13 8
Other research institutes 4 5 13
No private sector/other experience 16 9 12
Missing values - - 1
No. patents until 2005
(1st/2nd/3rd quartile)
1/2/8.5 1/2/5.75 1/2/6
No. patents 2006–2012
(1st/2nd/3rd quartile)
0/1/2.75 0/1/3 0/1/2
No. SSCI-E publications until 2005
(1st/2nd/3rd quartile)
0/2/12.25 0/1/11.5 13/29/60
No. SSCI-E publications 2006–2012
(1st/2nd/3rd quartile)
0.25/3/15 0.25/3/12.75 8/21.5/55
H. M. Grimm, J. Jaenicke
123
differences in the mean number of SSCI-E publications until 2005 between the Old and the
New Laender in our survey. The top quarter researcher in Thuringia or Saxony published
less than the lower quarter researcher in Bavaria. The median researcher in Thuringia
published two papers and the median researcher in Saxony only one paper in comparison
with 29 papers in Bavaria. These remarkable differences may be attributed to the trans-
formation process of the New Laender and a lack of publication habits in international
journals recorded in the SSCI-E. The difference decreases throughout the most recent
years. Nevertheless, the number of SSCI-E publications after 2005 differs significantly on
the 5 % level. Additional descriptive analysis reveals that around one-third of the
respondents were from the faculty of medicine (32 %), followed by engineering (25 %),
electrical engineering (12 %), biology (10 %), and physics (9 %). Other faculties are
represented only by a few respondents.
5 Granger-causality results: spillover between patent applications and publications
We estimate 14 different specifications of the dynamic panel model, estimated with cross
section SUR according to Eq. (1), to test the Granger-causal influence from patenting to
publications and 14 for the reverse direction including step by step group and Laender-
specific behavior and interaction effects. Ranking the different specifications of the
dynamic panel models by the Akaike information criterion AIC, we present the best and
the second best results in Table 3.
4
The best model explaining the number of publications including dummy variables for
Thuringia has significant interaction effects between lagged publication and patent output and
the age of the researcher (see Table 3, first column). The adjusted R
2
is rather high with 0.79.
The Granger-causality-v
2
-statistic measures the joint influence of the main and the interac-
tion effect and is equal to 5.961. This statistic is v
2
(2) distributed and accordingly, the Granger
causality test rejects the non-causality hypothesis on a significance level of 5.1 %. In this
specification, the main effect as well as the interaction effect between age and past number of
patent applications on publication is significant. The expected marginal effect of one addi-
tional patent application until 2005 on the publication record after 2005, i.e.
oE½ðNo:of publication after 2005Þ=oðNo:of patent applications until 2005Þ
¼2:745 0:049 age;
depends on the age of the patentee. One additional patent application has a positive impact
on future publication if the patentee is not older than 56 years. Evaluated at the mean age
of 44.7 years, one additional patent application increases future publications by 0.56. If the
patentee is older than 56 years—this holds for 20 % of our sample—we see an increasing
trade-off between patent application and publication. If we compute the marginal effect of
past patent applications on present patent applications we find a positive effect until the age
of 65, but the effect size decreases with increasing age. Additionally, other personal
characteristics have at least a weak significant impact on the publication record. Holding a
Habilitation, private or non-university working experience, and working at a public uni-
versity increases the number of publications. Working at the discipline of physics increases
the number of publications significantly in comparison with the reference discipline of
medicine.
4
All other results are available on request.
Crowding-out or reinforcement?
123
Table 3 Dynamic panel estimation results and granger-causality tests
Dependent variable No. of publications after 2005 No. of patent applications after 2005
Const. -25.180 34.202 7.18 6.928
[-1.797]* [2.296]** [3.335]*** [2.988]***
No. of publications until 2005 2.999 0.430 0.005 0.008
[9.077]*** [3.575]*** [0.686] [1.011]
No. of publications until 2005 9age -0.046
[-7.246]***
No. of publications until 2005 9Habilitation 0.427
[2.750]***
No. of publications until 2005 9Saxony 0.065 0.061
[2.608]** [2.419]**
No. of publications until 2005 9Thuringia 0.000
[0.008]
No. of patent applications until 2005 2.745 0.688 0.105 0.086
[2.431]** [1.809] [2.625]** [1.805]*
No. of patent applications until 2005 9age -0.049
[-2.426]**
No. of patent applications until 2005 9Habilitation -1.263
[-2.418]**
No. of patent applications until 2005 9Saxony 0.332 0.348
[3.428]*** [3.465]***
No. of patent applications until 2005 9Thuringia 0.025
[0.353]
Age 0.160 -0.880 -0.076 -0.083
[0.662] [-3.440]*** [-2.380]** [-2.353]**
Female 1.918 -3.913 -0.821 -0.752
[0.251] [-0.380] [-0.614] [-0.553]
H. M. Grimm, J. Jaenicke
123
Table 3 continued
Dependent variable No. of publications after 2005 No. of patent applications after 2005
No PhD -0.780 -10.896 -1.62 -1.833
[-0.164] [-1.809]* [-1.818]* [-1.880]*
Habilitation 8.656 5.180 -0.776 -0.750
[1.743]* [0.611] [-1.162] [-1.093]
Private or non-university 7.869 11.838 0.804 0.993
Work experience [1.860]* [2.160]** [1.259] [1.432]
University affiliation 13.316 7.274 -3.28 -3.131
[2.045]** [0.847] [-2.350]** [-2.094]**
Construction engineering 3.092 -17.067 0.04 -0.183
[0.330] [-1.258] [0.019] [-0.085]
Biology -5.169 -8.305 0.221 0.404
[-0.652] [-0.819] [0.242] [0.430]
Chemistry 9.150 -5.760 1.1 0.981
[1.174] [-0.568] [0.933] [0.804]
Electrical engineering 0.011 -16.270 2.44 2.488
[0.002] [-2.126]** [2.044]** [2.040]**
Computer science -8.976 -13.294 0.485 0.549
[-1.050] [-1.175] [0.312] [0.334]
Engineering 5.383 2.993 1.17 1.202
[0.999] [0.432] [1.227] [1.207]
Physics 17.824 8.595 -1.4 -1.227
[2.383]** [0.895] [-1.238] [-1.033]
Saxony -2.04 -1.693
[-1.991]* [-1.517]
Thuringia 4.168000 7.308 0.642
[1.083] [1.349] [0.646]
Crowding-out or reinforcement?
123
Table 3 continued
Dependent variable No. of publications after 2005 No. of patent applications after 2005
Observations 92 92 92 92
R-squared: 0.835 0.724 0.553 0.559
Adj. R-squared: 0.794 0.657 0.442 0.426
AIC: 8.606 9.117 5.04 5.090
F-statistic: 20.476 10.663 5.01 4.217
Granger-causality-v
2
-statistic 5.961* 6.0285** 8.012** 7.982**
Marginal probability 0.051 0.049 0.0182 0.046
The baseline researcher is male, holds a PhD, and works at the discipline of medicine
The t-statistics of the estimated coefficients are presented in squared brackets and * represents the 10 %, ** the 5 %, and *** the 1 % significance level. Granger-causality-
test results are presented in bold letters
H. M. Grimm, J. Jaenicke
123
The second best estimation results according to the AIC presented in Table 3, column 2
highlights the role of holding a Habilitation. According to the estimation results the
expected difference in the publication record with or without holding a Habilitation is
equal to
oE½ðNo:of publication after 2005Þjhabilitation¼1ðNo:of publication after 2005Þjhabilitation¼0Þ
¼0:427 No:of publications until 2005 1:263 No:of patent applications until 2005 þ5:180
and depends on the number of past publications and past patent application success.
Evaluated at the mean, the publication record is 8.4 publications higher when holding a
Habilitation in comparison to not holding a Habilitation. One additional patent application
decreases this difference by 1.26 publications.
Laender specific institutions play an important role predicting the patenting success.
Both models, chosen by the AIC, incorporate interactions between research outputs with
Laender dummy variables. Looking at the Granger-causal influence from the past publi-
cation record on the number of future patent applications (Table 3, column 3), we find that
test results are Laender-specific. The main effect is small and insignificant. Including the
interaction with the Laender dummy Saxony, the overall Granger-causality influence
became highly significant. If the patentee works in Saxony, the number of patent appli-
cations will be 1.6 applications higher in comparison to not working in Saxony. One
additional publication increases this difference by 0.065 patent applications. Further, age,
not yet holding a PhD, and the university affiliation affect the number of patent applica-
tions negatively; and working at the discipline of electrical engineering affects the number
positively in comparison to the baseline.
The last model includes the interactions with the dummy variable Saxony and addi-
tionally the interactions with Thuringia. Looking at the t-statistic, the interactions with
Thuringia are not separately significant. Nevertheless, the overall Granger-causal influence
running form publishing to patenting is highly significant.
Summarizing our Granger-causality results, we find a feedback relation between both
research activities at least for some groups. More publications will increase the number of
future patent applications in Saxony and more patent applications will increase the number
of future publications across all Laender if the patentee is not older than 56. Looking at the
mean age, we are not able to find a crowding out effect. With respect to the estimation of
patent activities, we find significant differences between the German Laender. Personal
characteristics, disciplines and academic degrees additionally play an important role
explaining patenting and publishing activities (Table 3).
6 Star scientists
The concept of star scientists was developed and applied by Zucker et al. (1998) who
presented a study finalized at the University of California, Los Angeles (UCLA), of elite
scientists in the field of biotechnology. They defined star scientists for their project as those
who had recorded more than 40 genetic-sequence discoveries or had authored at least 20
articles reporting such discoveries, until 1990. Their results indicated that the very best star
scientists play central roles in both the development of new knowledge and its successful
commercialization. Hess and Rothaermel (2011) defined star scientists by not only using
quantitative but also qualitative tools of measurement such as the number of citations
linked to publications as well as receiving a Nobel Prize. Both criteria definitely indicate
Crowding-out or reinforcement?
123
Table 4 Profiling six star scientists using number of patents until 2005 and number of publications after 2005
Patent i14 50 77 25 81 66
German Land Thuringia Saxony Bavaria Thuringia Bavaria Bavaria
Number of patents
until 2005
41 40 26 13 1 40
Number of patents
after 2005
4184 521
Number of SSCI-E
publications until
2005
0 7 96 124 164 65
Number of SSCI-E
publications after
2005
22 58 81 142 183 54
Age (in 2005,
categorized for
publication)
61–65 41–45 51–55 46–50 41–45 50–55
University Type University University University University University University
Faculty Engineering Engineering Engineering Chemistry Biology Physics
Status in 2012
(Highest university
degree at the time
of the survey)
Professor, retired
(Habilitation and Dr.
honoris causa mult.)
Professor
(PhD)
Professor (PhD) Professor
(Habilitation)
Professor
(Habilitation)
Professor (Habilitation)
Private or non-
university sector
working experience
No Yes Yes Yes No Yes
Motives for patent
application
Visibility and benefit of
invention; raising
university’s reputation
Intellectual
property
rights
protection
Intellectual property rights
protection; visibility and
benefit of invention
Raising own
and
university’s
reputation
Financial incentives;
visibility and benefit
of invention
Financial incentives;
entrepreneurial job
option; visibility and
benefit of invention
raising university’s
reputation
H. M. Grimm, J. Jaenicke
123
Table 4 continued
Patent i14 50 77 25 81 66
Transfer of
patent
No Yes Yes No Yes Yes
Law reform
supportive
No
(Likert =5)
Rather yes
(Likert =2)
Rather no
(Likert =4)
No
(Likert =5)
Neutral
(Likert =3)
Rather no
(Likert =4)
Open
answers
Academic
entrepreneurship
(using a patent
for start-up or
spin-off) is not
possible on your
own
More professional
specialized staff in
PEA and TTOS
(both Likert
scale =5) to
support academic
innovators
30 % rule no financial
incentive; better legal
support by specialized
lawyers; more professional
specialized staff in PEA and
TTOs (both Likert
scale =5)
30 % rule no financial
incentive; lack of big
industry in region;
decreased personal
involvement in
application process
negatively perceived
Lack of professional
specialized advice
and staff; PEA and
TTO not helpful
(both Likert
scale =5)
30 % rule no financial
incentive
Crowding-out or reinforcement?
123
that we are dealing with a star scientist. As a matter of fact, there was no Nobel Prize
winner amongst academic patentees in Germany who applied for a patent in 2005, in our
survey and beyond. In other words, it would have been difficult to transfer this concept to
our study. Further, we selected two East German and one West German Land for our
survey. Due to the specific East German academic history and context, we hardly find any
scientists who have a comparable publication history than their West German colleagues.
East German academics published in peer reviewed articles but hardly any of them were
published in journals which are listed in the Web of Science. In other words, East German
academics do not have a ‘‘publication history’’ or ‘‘citation history’’ as their West German
colleagues. Nonetheless, we found it specifically interesting to address the topic from a
unified perspective by considering the peculiarities of East and West German scientists and
their personal and institutional characteristics which might influence patenting and pub-
lishing activities. We have therefore selected star scientists according to the sheer number
of number of SSCI-E publications until and after 2005 as well as number of patents until
and after 2005 and looked at an efficient combination of both research activities in an
intertemporal perspective. This approach was very much influenced by the above-men-
tioned concept of star scientists developed and applied by Zucker and Darby (1996). We
have taken a more general and broader route of investigation for our study.
By graphical inspection, the data reveals a considerable degree of a star scientist effect
(c.f. Zucker et al. 1998). We used a two-dimensional approach to define so-called star
scientists in our research: We counted the number of patent applications in 2005 or earlier
and investigated the SSCI-E publication record after the patent application in 2005. The
question ‘‘did you apply for a patent before 2005’’ was raised in the questionnaire and
sufficiently answered by the respondents. The results are represented in Fig. 1, left panel.
Especially the performances of respondents 81, 25, and 77 are exceptional with regard to
other research records. Respondent 50 is also outstanding and has a few more publications
than respondent 66. Additionally, respondent 14 has one more patent application than
respondents 50 and 66 but obviously fewer publications. By looking at both research
dimensions, the star researchers 81, 25, 77, 50, and 14 are efficiently accomplishing both
activities. As displayed in Fig. 1, the increase of patent applications results in reduced
future publication activities. Additionally, we see Laender and academic qualification
specific differences.
We further profiled the star scientists by transferring information raised on age, private
sector working experience, transfer or commercialization of patent, assessment of insti-
tutional and other environmental conditions.
Two star scientists came from Thuringia, three from Bavaria and one from Saxony.
The most successful researcher with respect to patent applications, held multiple doctoral
degrees honoris causa, was over 60 years old in 2005, and had no SSCI-E publications
record in the past, he/she started a successful publication record in most recent years.
The other extraordinary star researcher registered successfully for one patent only (until
2005), is aged 41–45 years and published 183 (164) papers after (until) 2005. One
researcher from Bavaria applied for 41 patents until 2005 and published 65 refereed
papers before 2005 and 54 refereed papers after 2005. The star scientist successfully
started-up a company without any linkage with the employer university. Only one
researcher in our survey held a top ten position simultaneously in both categories. He or
she applied for 26 patents and published 96 papers after 2005. In 2012, all star scientists
held a Habilitation and were professors at a public university. Three out of six were
affiliated with the engineering faculty, one scientist is a physicist. The most successful
stars with regard to publications belonged to the faculty of chemistry and biology. Three
H. M. Grimm, J. Jaenicke
123
out of six star researchers rated the law reforms on a four or five point Likert scale to be
bad or very bad; and the support by TTO and PEA on five point Likert scale to be very
bad (Grimm and Jaenicke 2012). As hypothesized in Sect. 1, the findings confirm that
frequent patenting and publishing activity does imply that researchers aim at protecting
intellectual property rights, on the one hand, and reach for visibility of an invention, on
the other hand. Taking the entrepreneurial route is not a major motive of the star
scientists in the three selected German Laender with the exception of one researcher who
successfully started-up a company. Nonetheless, all researchers confirm that a major
motive to patent is associated with the intention to benefit from an invention. A con-
siderable number of patents (four out of six) have been successfully transferred to the
private sector. Academic excellence expressed as reinforcement of patenting and pub-
lishing is associated with a strong interest to knowledge transfer activities though aca-
demic entrepreneurship is not of major concern.
If we look at the spillover from the publication record until 2005 to the patent application
record after 2005 in Fig. 1, right panel, we see again some crowding-out effect. Scientific
stars that are extremely productive with respect to publications have only a few patent
applications in the future and vice versa. Two out of four stars presented in the right panel are
already known from the left panel. The new ones are respondent 83, between 56 and 60 years
old with 208 publication until 2005 and 40 publications after 2005; he/she has no Habilitation;
and the star scientist number 48 holding two doctoral degrees, a professorship and a pro-
fessorship honoris causa, who was member of German Council of Science and Humanities,
published over both time periods 140 articles and holds 50 patent applications (Table 4).
7 Conclusion
This paper investigates the relationship between patenting and publishing (and vice versa)
as well as the influence of personal characteristics and selected institutional assets of
0
40
80
120
160
200
0 10 20 30 40 50
No of patent applications until 2005
No of publications 2006-2012
81
25
77
66
50
14
0
5
10
15
20
25
30
0 40 80 120 160 200 240
No of publications until 2005
No of patent applications 2006-2012
48
25
81 83
Thuringia, no habilitation Thuringia, habilitation
Saxony, no habilitation Saxony, habilitation
Bavaria, no habilitation Bavaria, habilitation
Fig. 1 Identifying star scientists
Crowding-out or reinforcement?
123
academic patentees to pursue, align or specialize in one or both activities. With this
approach, we aim at filling a scientific gap as suggested by Czarnitzky et al. (2009)or
Stephan et al. (2007). We also employ the dataset to profile so-called star scientists to gain
firsthand information about their individual characteristics, motives and ambitions.
7.1 Factors influencing publishing
Empirical findings drawn from Granger-causality tests in a dynamic panel model using
survey data of university patentees matched with the SSCI-E data base show that patenting
activity has in general a positive and significant influence on scientific publishing. Our
findings confirm the above-presented relevant research about the positive patenting–pub-
lishing relationship within US academia (Stephan et al. 2007; Azoulay et al. 2009; Fabrizio
and Di Minin 2008); for Italian scientists (Breschi et al. 2008; Baldini et al. 2007); British,
French and Belgium academics (Crespi et al. 2011; Baldini et al. 2007; Carayol 2007; van
Looy et al. 2006); and finally German professors whose patents were assigned to uni-
versities (Czarnitzky et al. 2009). We control for different personal and research envi-
ronmental characteristics. Holding a Habilitation, working at a university, working in the
discipline of physics are significant factors increasing the publication record in the future.
Furthermore, we find that the seniority of the patentee, measured by the age of the
respondents, interacts significantly with the research output. This confirms our expectation
and is consistent with previous research that seniority, job stability and an advanced career,
as reflected by a Habilitation, positively influence patenting behavior (Moutinho et al.
2007; Carayol 2007). With increasing age (elder than 56), we notice a trade-off in the sense
that more patent activities will negatively affect the future publication record. Due to the
small number of respondents we are not able to address interactions between patenting and
different disciplines. By controlling for patenting and other personal and institutional
characteristics, we find that physicists have a higher publication output in comparison to
the reference discipline medicine. Laender differences do not play an important role in
context of the publication activities of the respondents in the final estimation equation.
Evaluated for the mean of age, one additional publication in the past, or one additional
patent application in the past increases the number of future publications, though this
increase decreases as researchers advance in years. For patentees older than 56 years, the
effect of patenting on publishing turns negative.
7.2 Factors influencing patenting
In several specifications, we find Granger-causal effects from publishing to patenting.
Interactions with the Laender specific research environment play an important role. The
final model, chosen by the AIC, shows a positive impact of publication on patenting for
patentees working in Saxony. Under the assumption of identical personal characteristics,
scientists working at universities of applied sciences have a higher patent application
record in the period 2006–2012 in comparison with their colleagues from universities.
Looking at the subsample of university researchers in Bavaria, the past publication
success seems to be the most influential variable, as revealed by additional estimation
results for this subsample. Successful patent applications have a strong positive influence
on other research activities in Saxony. In our survey, the most productive patent applicants
came from the New Laender.
The academic status plays an important role in some specifications. Additionally, we
find that non-university working experience does not have a significant influence on
H. M. Grimm, J. Jaenicke
123
patenting. This may be a result of the small number of observations. Most relevant research
underlines that involvement in industry-oriented activities is associated with the scientist
´s
motivation to patent (for example, Moutinho et al. 2007; Agrawal and Henderson 2002;
Crespi et al. 2011). Discipline-wise, scientists from the field of electrical engineering have
a higher number of patent applications in comparison with medical scientists.
7.3 Profile of star scientists
If we look at the spillover from the publication record until 2005 to the patent application
record after 2005, we see a strong crowding-out effect in the sense that scientific stars who
are extremely productive with respect to publications only have few patent applications in
the future and vice versa. In conjunction with the descriptive statistics and profile of
selected star scientists we conclude that scientific excellence defined as frequent patenting
and publishing is associated with the motive to benefit from an invention although the star
scientists do not emphasize that academic entrepreneurship is a significant career option.
This insight may have major consequences for academic policy and new forms of aca-
demic management because the current policy aims at aligning patenting and entrepre-
neurial activity in academia. Further emphasis should be directed to link star scientists and
academic inventors with non-university actors and/or students who may exploit a patent for
commercial use. Overall, our results confirm that motives, skills, and status are associated
with entrepreneurial behavior or at least with the vested interested to transfer innovative
knowledge from academia into practice. These results are important for fine-tuning current
policies to promote academic patenting and entrepreneurship by creating diverse forms of
knowledge exchange with various actors as well as intensified university-business
interaction.
7.4 Research perspectives
Further research should be envisaged with regard to the following initial interpretations of
our findings: the successful patent application contributed positively to the future publi-
cation output of researchers only up to a certain age (56).
Despite its contributions, our study has some limitations. Our results require further
corroboration given the design and treatment of the database used for our analysis. First,
although the response rates of the surveys in Thuringia and Saxony were high (around
45 %), the response rate in Bavaria was lower (27.2 %) due to diverse reasons such as a
very low rate of patent applications by university academics (46 in total or 0.34 % of all
patent applications in Bavaria according to DPMA 2006, p. 13) and a high number of team
applications and the reluctance of all team members to answer to our questionnaire but also
time constraints etc. In total, 92 questionnaires were answered and returned to us, which
serves rather as a small-scale sample for a dynamic panel model. Second, the starting
sample may be biased because we selected academic patentees for our survey, in other
words, those scientists who successfully went through a patent application process. We
addressed this group of academics on purpose assuming a) that they are knowledgeable to
evaluate new public policy aiming at simplifying and promoting patent applications at
universities; b) that they pursue both activities (patenting and publishing) as career goals.
We considered adding a control group (non-patenting academics), but the development of a
dataset constructed from the information of our survey plus researcher’s curriculum vitae
(CV) turned out to be very time consuming. We also observed that CVs of academic
patentees are rather accessible and available online than those of non-patenting colleagues,
Crowding-out or reinforcement?
123
which explains our approach. Nonetheless, there could be a bias in the sense that aca-
demics who successfully patent and publish are very much career-oriented which may
influence the effect of prior patenting on future publishing. Also, extracting comparable
information from CVs was a challenge; CVs hardly refer, for example, to information
about collaboration with industry, and this issue was only hesitantly or too vaguely
answered in the survey. Further research could address these limitations by extending the
sample size to increase the potentials for drawing generalizations and by using our research
to fine-tune potential questions to be included in a survey. A larger sample is in any case
needed to validate our results. Further, by including a higher number of German Laender
for comparative studies, the depth of information on personal characteristics and institu-
tional assets of scientists to promote and align patenting and publishing at German uni-
versities will improve. Nevertheless, this is the first attempt to develop a comprehensive
dataset of German scientists with the goal to contribute to a field of research which needs
far more attention to design useful public policy.
Acknowledgments The research project was initially supported by the German Ministry of Economics
(ERP Transatlantic Program). We are grateful for having had the opportunity to present an early version of
this paper at the workshop ‘‘Academic Policy and the Knowledge Theory of Entrepreneurship’’ organized by
Zoltan Acs, David B. Audretsch and Erik E. Lehmann at the University of Augsburg, Germany, in August
2012. We thank the discussant, Alexander Dilger, and session participants for valuable comments. Further,
we acknowledge that the PATON | Landespatentzentrum Thu¨ringen provided us with data on patent
applications in 2005 as well as the survey respondents for their commitment.
References
Agrawal, A., & Henderson, R. (2002). Putting patents in context: Exploring knowledge transfer from MIT.
Management Science, 48, 44–60.
Audretsch, D. B. (2013). From the entrepreneurial university to the university for the entrepreneurial
society. Journal of Technology Transfer. doi:10.1007/s10961-012-9288-1.
Audretsch, D., & Aldridge, T. (2009). Scientist commercialization as conduit of knowledge spillovers. The
Annals of Regional Science, 43(4), 897–905.
Azoulay, P., Ding, W., & Stuart, T. (2007). The determinants of faculty patenting behavior: Demographics
or opportunities? Journal of Economic Bevahior & Organization, 63(4), 599–623.
Azoulay, P., Ding, W., & Stuart, T. E. (2009). The impact of academic patenting on the rate, quality, and
direction of (public) research output. Journal of Industrial Economics, 57(4), 637–676.
Balconi, M., Breschi, S., & Lissoni, F. (2004). Networks of innovators and the role of academia: An
exploration of Italian patent data. Research Policy, 33(1), 127–145.
Baldini, N., Grimaldi, R., & Sobrero, M. (2007). To patent or not to patent? A survey of Italian inventors on
motivations, incentives, and obstacles to university patenting. Scientometrics, 70(2), 333–354.
Bercovitz, J., & Feldman, M. (2008). Academic entrepreneurs: Organizational change at the individual
level. Organization Science, 19(1), 69–89.
Bishop, K., D’Este, P., & Neely, A. (2011). Gaining from interactions with universities: Multiple methods
for nurturing absorptive capacity. Research Policy, 40(1), 30–40.
Blumenthal, D., Campbell, E. G., Anderson, M. S., Causino, N., & Louis, K. S. (1996). Withholding
research results in academic life science: Evidence from a national survey of faculty. Journal of the
American Medical Association, 277(15), 1224–1228.
BMBF (Bundesministerium fu
¨r Bildung und Forschung). (2001). 1. Fo
¨rderrichtlinie des Bundesministeri-
ums fu
¨r Bildung und Forschung zur BMBF-Verwertungsoffensive—Verwertungsfo
¨rderung—vom
27.07.2001. Bundesanzeiger, 144/4.8.2001, 16657.
BMBF (Bundesministerium fu
¨r Bildung und Forschung). (2002). Zur Einfu
¨hrung der Neuheitsschonfrist im
Patentrecht—ein USA-Deutschland-Vergleich bezogen auf den Hochschulbereich. Schlussbericht.
Bonn.
Breschi, S., Lissoni, F., & Montobbio, F. (2008). University patenting and scientific productivity: A
quantitative study of Italian academic inventors. European Management Review, 5(2), 91–109.
H. M. Grimm, J. Jaenicke
123
Buenstorf, G. (2009). Is commercialization good or bad for science? Individual-level evidence from the
MaxPlanck Society. Research Policy, 38(2), 281–292.
Calderini, M., Franzoni, C., & Vezzulli, A. (2007). If star scientists do patent: The effect of productivity,
basicness and impact on the decision to patent in the academic world. Research Policy, 36(3),
303–319.
Campbell, E. G., Carridge, B. R., Gokhale, M., Berenhaum, L., Hilgartner, S., Holtzman, N. A., et al.
(2002). Data withholding in academic genetics. Journal of the American Medical Association, 287(4),
473–480.
Carayol, N. (2007). Academic incentives, research organization and patenting at a large French university.
Economics of Innovation and New Technology, 16(2), 119–138.
Crespi, G., D’Este, P., Fontana, R., & Geuna, A. (2011). The impact of academic patenting on university
research and its transfer. Research Policy, 40(1), 55–86.
Czarnitzky, D., Gla
¨nzel, W., & Hussinger, K. (2009). Heterogeneity of patenting activity and its implica-
tions for scientific research. Research Policy, 38(1), 26–34.
D’Este, P., Mahdi, S., Neely, A., & Rentocchini, F. (2012). Inventors and entrepreneurs in academia: What
types of skills and experience matter? Technovation, 32(5), 293–303.
DPMA (Deutsches Patent- und Markenamt). (2006). Jahresbericht 2006.Mu
¨nchen: Deutsches Patent- und
Markenamt.
Dumitrescu, E. I., & Hurlin, C. (2012). Testing for Granger non-causality in heterogeneous panels. Eco-
nomic Modelling, 29(4), 1450–1460.
Fabrizio, K. R., & Di Minin, A. (2008). Commercializing the laboratory: Faculty patenting and the open
science environment. Research Policy, 35(6), 790–807.
Geuna, A. (2001). The changing rationale for European university research funding: Are there negative
unintended consequences? Journal of Economic Issues, 32(3), 607–632.
Geuna, A., & Mowery, A. (2007). Publishing and patenting in US and European universities. Economics of
Innovation and New Technology, 16(2), 67–70.
Geuna, A., & Nesta, L. (2006). University patenting and its effects on academic research: The emerging
European evidence. Research Policy, 35(6), 790–807.
Grandi, A., & Grimaldi, R. (2005). Academics organizational characteristics and the generation of suc-
cessful business ideas. Journal of Business Venturing, 20, 821–845.
Granger, C. W. J. (1969). Investigating causal relations by econometric models and cross-spectral methods.
Econometrica, 37(3), 424–438.
Greif, S., & Schmiedl, D. (2006). Patentatlas Deutschland—Ausgabe 2006.Regionaldaten und Er-
findungsta¨tigkeit. Mu
¨nchen: Deutsches Patent- und Markenamt.
Grimm, H. M., & Jaenicke, J. (2012). What drives patenting and commercialization activity at East German
universities? The role of new public policy, institutional environment and individual prior knowledge.
Journal of Technology Transfer, 37(4), 454–477.
Gulbrandsen, M. (2005). But Peter’s in it for the money. The liminality of entrepreneurial scientists. VEST
Journal for Science and Technology Studies, 18(1–2), 49–75.
Hess, A. M., & Rothaermel, F. T. (2011). When are assets complementary? Star scientists, strategic alliances
and innovation in the pharmaceutical industry. Strategic Management Journal, 32(8), 895–909.
Hoye, K., & Pries, F. (2009). ‘Repeat commercializers’, the ‘habitual entrepreneurs’ of university–industry
technology transfer. Technovation, 29(10), 682–689.
Hu
¨lsbeck, M., Lehmann, E., & Starnecker, A. (2013). Performance of technology transfer offices in Ger-
many. Journal of Technology Transfer, 38(3), 199–215.
Hurlin, C. (2005). Un test simple de l‘hypothe
`se dans un mode
`le de panel he
´te
´roge
`ne. Revue e´conomique,
56, 799–809.
Kim, J. W. (2008). University patenting and scientific productivity. European Management Review, 5(2),
111–113.
Klitkou, A., & Gulbrandsen, M. (2009). The relationship between academic patenting and scientific pub-
lishing in Norway. Scientometrics, 82, 93–108.
Krimsky, J. (2003). Small gifts, conflicts of interest, and the zero-tolerance threshold in medicine. The
American Journal of Bioethics, 3(3), 50–52.
Landry, R., Amara, N., & Quimet, M. (2007). Determinants of knowledge transfer: evidence from the
Canadian researchers in the natural sciences and engineering. Journal of Technology Transfer, 32(6),
561–592.
Landry, R., Saihi, M., Amara, N., & Quimet, M. (2010). Evidence on how academics manage their portfolio
of knowledge transfer activities. Research Policy, 39(10), 1387–1403.
Crowding-out or reinforcement?
123
Lowe, R., & Gonzalez-Brambilla, C. (2005). Faculty entrepreneurs and research productivity: Are faculty
entrepreneurs stars and is entrepreneurship a distraction? Presentation at the technology transfer
conference, 29 September 2005.
Meyer, M. (2006). Are patenting scientists the better scholars? An exploratory comparison of inventor-
authors with their non-inventing peers in nano-science and technology. Research Policy, 35(10),
1646–1662.
Meyer-Krahmer, F., & Schmoch, U. (1998). Science-based technologies: University-industry interactions in
four fields. Research Policy, 27, 835–851.
Moutinho, P., Fontes, M., & Godinho, M. (2007). Do individual factors matter? A survey of scientists’
patenting in Portuguese public research organisations. Scientometrics, 70(2), 355–377.
Mowery, D. C., & Sampat, B. V. (2005). The Bayh-Dole Act of 1980 and university-industry technology
transfer: A model for other OECD governments? Journal of Technology Transfer, 30(1/2), 115–127.
Murray, F., & Stern, S. (2007). Do formal intellectual property rights hinder the free flow of scientific
knowledge? An empirical test of the anti-commons hypothesis. Journal of Economic Behavior &
Organization, 63(4), 648–687.
Powers, J. B., & McDougall, P. P. (2005). University start-up formation and technology licensing with firms
that go public: A resource-based view of academic entrepreneurship. Journal of Business Venturing,
20(3), 291–311.
Rothaermel, F. T., Agung, S. S., & Jiang, L. (2007). University entrepreneurship: A taxonomy of the
literature. Industrial and Corporate Change, 16(4), 691–791.
Satterthwaite, F. E. (1946). An approximate distribution of estimates of variance components. Biometrics
Bulletin, 2(6), 110–114.
Shane, S. (2002). Selling university technology: Patterns from MIT. Management Science, 48(1), 122–137.
Shin, J. C., & Cummings, W. K. (2010). Multilevel analysis of academic publishing across disciplines:
Research preference, collaboration, and time on research. Scientometrics, 85, 581–594.
Stephan, P. E., Gurmu, S., Sumell, A. J., & Black, G. (2007). Who
´s patenting in the university? Evidence
from the survey of doctorate recipients. Economics of Innovation and New Technology, 16(2), 71–99.
van Looy, B., Callaert, J., & Debackere, K. (2006). Publication and patent behavior of academic researchers:
Conflicting, reinforcing or merely co-existing? Research Policy, 35(4), 596–608.
Walter, T., Ihl, C., Mauer, R., & Brettel, M. (2013). Grace, gold, or glory? Exploring incentives for
invention disclosure in the university context. Journal of Technology Transfer. doi:10.1007/s10961-
013-9303-1.
Welch, B. L. (1951). On the comparison of several mean values: An alternative approach. Biometrika, 38,
330–336.
Welter, F. (2011). Contextualizing entrepreneurship—conceptual challenges and ways forward. Entrepre-
neurshipTheory and Practice, 35(1), 165–178.
Wright, M. (2013). Academic entrepreneurship, technology transfer and society: Where next? Journal of
Technology Transfer. doi:10.1007/s10961-012-9286-3.
Wright, M., Birley, S., & Mosey, S. (2004). Entrepreneurship and university technology transfer. Journal of
Technology Transfer, 29(3–4), 235–246.
Zucker, L. G., & Darby, M. R. (1996). Star scientists and institutional transformation: Patterns of invention
and innovation in the formation of the biotechnology industry. Proceedings of the national academy of
sciences, November 12, 93(23), 12709–12716.
Zucker, L. G., Darby, M. R., & Brewer, M. B. (1998). Intellectual human capital and the birth of U.S.
biotechnology enterprises. The American Economic Review, 88(1), 290–306.
H. M. Grimm, J. Jaenicke
123
