Research Policy 51 (2022) 104409

Contents lists available at ScienceDirect

Research Policy
journal homepage: www.elsevier.com/locate/respol

Is open innovation imprinted on new ventures? The cooperation-inhibiting

legacy of authoritarian regimes
Michael Wyrwich *, Philip J. Steinberg , Florian Noseleit , Pedro de Faria
Faculty of Economics and Business, Department of Innovation Management and Strategy, University of Groningen, Nettelbosje 2, Groningen, AE 9747, the Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: Extant research explores the role played by individuals and, in particular, founders in defining open innovation
Open innovation strategies at the firm level. We join this discussion by combining insights from imprinting literature that explores
Imprinting theory the enduring impact of a founder’s personal history, with inputs from literature that stresses the impact of past
Authoritarian regimes
experience on trust formation. We suggest that founders are less likely to engage in open innovation if their
R&D cooperation
Surveillance
experiences engender a generalized lack of trust. We use a unique database that includes East and West German
Organizational decision-making founders to identify regional differences in activities conducted by authoritarian regimes that could inhibit trust.
We find that founders who were exposed to high levels of secret police surveillance in the former socialist
German Democratic Republic (GDR) are less likely to engage in interfirm R&D cooperation. We contribute to the
literature on open innovation by exploring how a founder’s social, political, and cultural backgrounds influence
strategic decisions related to open innovation, and to recent imprinting literature by showing that variation in
oppressive enforcement practices in authoritarian regimes, such as surveillance activities, can leave an enduring
imprint. Our findings complement recent insights on ideological imprinting effects on young firms’ decision-
making.

1. Introduction and conceptual background
“The net of spies was really broad. In nearly every institution, even in the
churches […], there were many of them. I remember telling a leading Stasi
(secret police) officer, ’If you had sent an informant to me, I would surely
have recognized him.’ His-answer was, ‘We didn’t send anyone. We took
those who were around you.’" Hubertus Knabe (former director of the Stasi
prison memorial) (Knabe, 2014).
In the last decades, closed innovation systems characterized by firms
conducting most R&D activities internally, are transitioning to open
innovation systems in which firms manage knowledge flows and search
for knowledge sources with external actors outside their organizational
boundaries (cf., Chesbrough and Bogers 2014, Chesbrough 2003, 2006).
This transition triggered researchers in innovation studies to explore the
open innovation phenomenon (cf., Dahlander and Gann, 2010, West
et al. 2014). They have mainly focused their attention to the identifi­
cation of open innovation’s benefits and risks (cf., Dahlander and Gann
2010, Ebersberger et al. 2021, Srinivasan et al. 2021), and to the un­
derstanding of why certain firms embrace open innovation and others do
not (Lee et al., 2010). While such discussions are often rooted in an
industry and/or organizational level (e.g., de Faria et al. 2020, 2010,
Keupp and Gassmann 2009), some scholars have recently shifted their
attention toward the micro-foundations of open innovation (Bogers
et al., 2018), and explore the role that individuals and, in particular,
founders play in leading an organization to adopt an open innovation
strategy (Classen et al., 2012; Eftekhari and Bogers, 2015). Hahn et al.
(2019), for instance, provide evidence that the professional background
of founders, specifically the presence of scientists in the founder team,
often leads to imprinting certain routines that encourage open innova­
tion practices.
Even though existing research has increased our understanding of
how founders’ human and social capital and their professional experi­
ence determine open innovation strategies, we know very little about if
and how founders’ social, political, and cultural backgrounds also play a
role regarding open innovation decision making. We build on research
that links a founder’s characteristics to open innovation and focus on
how a founder’s socioeconomic environment affects the firm’s likeli­
hood to adopt open innovation strategies. More specifically, we inves­
tigate how socialization in an authoritarian regime imprints a founder’s
later strategic decision-making at the firm level in a free-market

* Corresponding author.
E-mail addresses: m.wyrwich@rug.nl (M. Wyrwich), p.j.steinberg@rug.nl (P.J. Steinberg), f.noseleit@rug.nl (F. Noseleit), p.m.m.de.faria@rug.nl (P. de Faria).

https://doi.org/10.1016/j.respol.2021.104409
Received 9 February 2021; Received in revised form 11 October 2021; Accepted 18 October 2021
Available online 3 November 2021
0048-7333/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M. Wyrwich et al.
economy.
As illustrated by the quote at the beginning of the introduction,
authoritarian regimes actively limit individual freedom and the free
exchanges of ideas by implementing comprehensive surveillance activ­
ities. The purpose of these activities is to control and shape individuals’
views about how society and the economy should be organized (Svolik,
2012; Gerschewski, 2013). For citizens living under such regimes, these
surveillance activities create low trust environments where individuals
view each other with suspicion (Lichter et al., 2021).
In the second half of the 20th century, several South American (e.g.,
Chile, Argentina, and Brazil), European (e.g., Spain, Czech Republic, and
East Germany), Asian (e.g., South Korea), and African (e.g., South Af­
rica) countries transitioned from authoritarian regimes with closed
economies to democracies with market economies. Since the contrast
between systems is stark, these transitions led to economic and social
challenges that attracted the attention of economics and management
scholars (e.g., Aslund and Djankov 2014, Svolik 2012). These scholars
investigated the interaction between long-term imprinting caused by
authoritarian regimes’ efforts to control individual behavior, and an
individual’s behavior as an economic actor even after transitioning to
more open societies (Alesina and Fuchs-Schuendeln, 2007; Rainer and
Siedler, 2009).
Marquis and Qiao (2020) focus on how exposure to authoritarian
regimes imprints young firms’ internationalization decisions, and their
evidence begs for further research on authoritarian regimes’ imprint on
founders’ firm-level strategic choices. We integrate the insights from the
research on authoritarian regimes’ imprinting into the literature on
open innovation and argue that an authoritarian regime’s surveillance is
a trust-destroying mechanism that negatively affects a founder’s later
decision to participate in open innovation activities. Our assertion is
based on imprinting theory, which emphasizes the lasting influence of
historical experiences on individual and organizational behaviors and
outcomes (Stinchcombe, 1965; Marquis and Tilcsik, 2013; Simsek et al.,
2015; Ferriani et al., 2020). Our argument relies on two related ideas.
First, that a founder’s individual experiences, skills, and personal
characteristics influence subsequent strategic choices (Burton and
Beckman, 2007; Hahn et al., 2019; Bryant, 2014; Roche et al., 2020).
Second, that an individual’s socioeconomic context shapes these char­
acteristics (Tabellini, 2010). Bringing these two ideas together, we claim
that a founder’s strategic choices are shaped by her socioeconomic
context (Waguespack et al., 2018). Indeed, empirical research shows
that growing up in a specific political regime can leave a long-lasting
imprint on an individual (e.g., Alesina and Fuchs-Schuendeln 2007,
Lichter et al. 2021). Based on the evidence that not all individuals living
under an authoritarian regime have similar experiences (e.g., Lichter
et al. 2021), we claim that the imprinting of individuals stemming from
being exposed to the socioeconomic context of a particular authoritarian
regime is not uniform. We further argue that imprinting stemming from
differences in exposure to oppressive regime enforcement practices is a
mechanism that inhibits an individual’s ability to develop generalized
trust.
Trust can be defined as “a more or less well-grounded expectation
about the preferences of other people” (Delhey and Newton, 2005, 311),
and is considered to be a central determinant of social cooperation (e.g.,
Hardin 2006, Mayer et al. 1995). Higher levels of trust lead to a greater
willingness to accept the vulnerabilities of a third party (Ertug et al.,
2013). While there are different classifications of trust, we focus on
trusting strangers (Rotter, 1971; Dinesen and Bekkers, 2017), referred to
as generalized or social trust (Kramer and Lewicki, 2010; Nannestad,
2008). It is not entirely clear when and/or how individuals form
generalized trust levels, however, recent research by Glanville and
Paxton (2007) supports the social learning perspective that suggests that
generalized trust is formed by a variety of social interactions (Burns
et al., 2003; Rotter 1971). Although generalized trust varies across
cultures and countries (Ferrin and Gillespie, 2010; Huff and Kelley,
2003), it has been shown that individuals living in a communist system
2
Research Policy 51 (2022) 104409
tend to have lower levels of generalized trust (Bjørnskov, 2006; Rainer
and Siedler, 2009). Our conjecture that oppressive surveillance activ­
ities conducted by authoritarian regimes negatively affect generalized
trust levels is based on the work of Hardin (2006), who outlines that
trustworthiness gives rise to trust. Consequently, the oppressive
behavior of regimes negatively affects the general level of trustworthi­
ness (Paldam and Svendsen, 2001), and lowers an individual’s gener­
alized trust levels. As individuals’ trust levels typically remain constant
over time (Algan and Cahuc, 2010), this adverse effect has long-run
consequences and lower trust levels remain even when an individual
no longer lives under an oppressive regime (Bjørnskov, 2006; Rainer and
Siedler, 2009).
These enduring consequences also play a central role in the
conceptualization of imprinting that considers historical factors
(Marquis and Tilcsik, 2013; Simsek et al., 2015). To explore the effect of
imprinting in the context of open innovation strategies, we focus on an
open innovation strategy where trust is essential: the decision to engage
in R&D cooperation with other organizations. From all the different
types of open innovation (for an overview, see Dahlander and Gann
2010), inter-organizational cooperation is specifically dependent on
trust and cooperative behavior (Nooteboom et al., 1997; Faems et al.,
2008; Ertug et al., 2013). Due to the high degree of endogenous project
uncertainty and a high likelihood for incomplete contract protection,
opportunistic behavior in R&D cooperation partnerships is difficult to
avoid via hierarchical or markets mechanisms (Mayer and Nickerson,
2005; Ulset, 1996). Trust, therefore, plays an important role in the de­
cision to engage in R&D cooperation activities because it increases a
firm’s willingness to accept vulnerability towards a third party (Ertug
et al., 2013).
While relational trust between cooperating partners is critical (e.g.,
Krishnan et al. 2006, Lee and Cavusgil 2006), a more generalized trust
level within an organization is likely to affect its decision to initiate
cooperative activities with other organizations. For example, Ferrin and
Gillespie (2010) show that individuals from countries with high gener­
alized trust levels will be more cooperative than individuals from
low-trust countries. Because trust occurs between individuals rather
than organizations (Zaheer et al., 1998), the personal experiences of
central actors can have a substantial influence on how an organization’s
cooperative behavior evolves. Ertug et al. (2013) show that high levels
of generalized trust within a society positively influence firms’ perceived
trustworthiness of specific partners. Thus, we expect that erosion of
generalized trust levels among an organization’s founding members is
likely to impede the adoption of open innovation and cooperative R&D
strategies.
We address recent calls for a better understanding of the micro-
foundations of open innovation by combining the primary notion of
imprinting literature that prior history can have an enduring impact on
strategic choices with the insights from literature of generalized trust
that emphasizes the importance of past experiences on trust formation.
Our analysis makes two important contributions. First, building on
recent research on imprinting mechanisms, we offer a more fine-grained
analysis of the context and improve our understanding of the imprinting
environment (Simsek et al., 2015; Marquis and Qiao, 2020). Marquis
and Qiao (2020), for example, focus on the (uniform) ideological
imprinting of authoritarian regimes. Ideological imprinting refers to the
process of imprinting a “perceptual framework of interrelated ideas and
beliefs” (Wang et al., 2019, p. 1113). In contrast to Marquis and Qiao
(2020), we shed light on imprinting caused by an oppressive regime’s
enforcement practices. We empirically separate the effect of uniform
ideological imprinting from the effect of variation in exposure to an
authoritarian regime’s oppressive actions. We exploit an empirical
setting that is particularly pertinent to understand the imprinting effects
of authoritarian regimes’ surveillance practices on R&D cooperation. By
so doing, we expand our understanding of the micro-foundations of open
innovation in new ventures by highlighting the role of a founder’s prior
experiences in strategic decisions. We expect that those founders
M. Wyrwich et al.
experiencing more severe levels of surveillance and control will have
lower levels of trust that affect their decision to adopt open innovation
strategies. Second, we give a more nuanced perspective to the notion
that trust plays a central role in open innovation and R&D collaboration
(Carson et al., 2003; Brockman et al., 2018). Focusing on generalized
trust, we complement research on how relational trust impacts the
performance of collaborations (Faems et al., 2008), by suggesting that
generalized trust may be of crucial importance, specifically in the
context of new ventures.
2. Empirical approach
2.1. Study context
We study founders in East and West Germany about three decades
after German reunification to test our prediction that firms founded by
individuals who experienced high surveillance levels during the socialist
era are imprinted with low levels of trust and will be less likely to
cooperate with other firms. Before reunification, East Germany, the
former German Democratic Republic (GDR), was part of the Soviet bloc
and had a rigid authoritarian regime. In 1990, East Germany adopted
the formal institutional market economy framework of West Germany.
The opportunities for and the relative pay-offs of entrepreneurship
changed dramatically as a result of the transition. Entrepreneurship
usually plays a minor role in non-democratic authoritarian regimes
(Audretsch and Moog, 2020). The socialist institutions in the GDR were
particularly hostile toward entrepreneurial activities, and running a
privately owned venture in that environment was almost impossible
(Fritsch et al., 2014).
Like other authoritarian regimes, the GDR government had powerful
intelligence agencies and used undercover informants to surveil its cit­
izens. This created an environment of suspicion and long-lasting low
levels of trust, not only towards strangers but even towards friends and
family members (Bjørnskov, 2006; Rainer and Siedler, 2009). After the
fall of the Berlin Wall in November 1989, the East German intelligence
agencies were dissolved in preparation for German reunification in
October 1990. This historical reality means that East German founders
were exposed to an authoritarian regime, while West German founders
represent non-treated counterfactual observations.
One advantage of our data and the empirical set-up is that both allow
us to identify regional differences in surveillance levels and go beyond a
simple binary East-West comparison that would only capture a general
“GDR/East German effect”. Because there were, in fact, regional varia­
tions in GDR surveillance levels conducted by the East German Secret
Police (Stasi), we can test if these variations left varying imprints on the
founders.
The Stasi developed a system of recruiting undercover informants to
spy on their friends, neighbors, and families, which implied: “the
breakdown of the bonds of trust between officers and men, lawyers and cli­
ents, doctors and patients, teachers and students, pastors and their commu­
nities, friends and neighbors, family members and even lovers” (Childs and
Popplewell, 1996, p. 111). Lichter et al. (2021) find that East Germans
who were socialized in regions with high levels of surveillance reveal a
lower level of trust, even several years after reunification. We use the
same regional dataset to measure local spying intensity and analyze
whether growing up in spying-intense regions reduces the likelihood of
the respective East German founders to engage in R&D cooperation. If
this is the case, this would be in line with our prediction that the
trust-inhibiting effects of surveillance within authoritarian regimes
inhibit R&D cooperation. We include West German founders with strong
similarities in culture and language that were not subject to an author­
itarian regime’s influence as a benchmark for cooperative behavior.
These founders experienced a local environment where spying activity
was virtually non-existing.
3
Research Policy 51 (2022) 104409
2.2. Sample
Our analysis uses data from the GAW-ADIAB survey (also used in
previous research, e.g., Bublitz et al. 2018), a representative survey of
East and West German founders across six sample regions (for details,
see Wyrwich et al. 2018). The sample is based on the addresses of new
firms in manufacturing and knowledge-intensive business services
(KIBS) and is taken from the Establishment History Panel (BHP)
collected by the German Federal Employment Agency’s Institute for
Employment Research. The BHP includes information on all German
firms with at least one employee. All new manufacturing firms in the
sample regions that appeared in the BHP between 2003 and 2007 for the
first time1 and were still active in 2010 were contacted for the survey.
For KIBS, about 75–80% of the respective new start-ups were contacted.
The overall response rate was 31%, which is more than twice as large as
in the Global Entrepreneurship Monitor adult population survey, a
standard dataset for the analysis of entrepreneurship (GEM, 2020). The
GAW-ADIAB sample of technology-oriented manufacturing and
knowledge-intensive business service firms (based on the NACE sector
classification) comprises 772 observations.2
In order to ensure that the focal founder was socialized in the GDR
before German reunification and the dissolution of the Stasi we used two
additional criteria to create our final sample. First, because the GDR did
not exist before 1949, we excluded firms whose focal founder was born
before 1949. Second, we follow Lichter et al. (2021) and other studies on
legacy effects of the GDR (Fuchs-Schuendeln and Masella, 2016) that
utilized the same spying data, and considered only respondents born
before 1974, and therefore were at least 17 years old at the time of
German reunification in 1990, to ensure a clean treatment effect of being
exposed to Stasi surveillance activity. Thus, all respondents reached or
passed already the age (i.e., 18 to 25 years) that is regarded as crucial for
the formation of beliefs in accordance to social-psychology literature
and confirmative recent empirical evidence (for an overview, see Giu­
liano and Spilimbergo 2014). After data cleaning, our final sample in­
cludes 620 firms, with 214 firms involved in R&D activities. In a
robustness check, we consider founders born after 1973, and this full
sample comprises 714 firms with 251 firms engaging in R&D.
2.3. Variables
2.3.1. Dependent variable
Our dependent variable captures a firm’s involvement in interfirm
R&D cooperation activities. It is measured by a dichotomous variable and
refers to any R&D cooperation since the start-up’s foundation. The
survey question explicitly refers to R&D cooperation with other private
firms, just as in the standard EU community innovation survey (CIS,
2020).3
2.3.2. Variable of interest
Also in line with Lichter et al. (2021), we measure spying intensity as
the local share of the population working either as an official employee
or undercover informant for the Stasi in the region (at NUTS3 level)
1
Since the first appearance in the BHP indicates that the venture hired an
employee liable to Social Insurance for the first time, the actual start-up deci­
sion could have taken place several years earlier.
2
This comprises all high, medium-high, and medium-low technology-inten­
sive manufacturing industries and all knowledge-intensive service (NACE Rev
2, 2-digit level: 19-30, 33, 58-63, 69-75; Eurostat, 2018).
3
For description of the dataset, see https://ec.europa.eu/eurostat/web/mi
crodata/community-innovation-survey
M. Wyrwich et al.
Table 1
Definition of variables.
Name Description
Interfirm R&D Dummy variable that indicates
Cooperation whether the founder’s firm
cooperates on research and
development with other companies.
Spying intensity Local share of population (in%)
working either as an official
employee or undercover informant
for the secret police in the region
where the interviewed founder grew
up.
GDR origin Dummy variable that indicates
whether the founder lived in East
Germany in 1989.
Age Age of the founder at the time of the
interview.
Women Dummy variable that indicates
whether the founder is a woman.
Start-up size Number of employees in the year in
which the first hires took place.
Start-up period Dummy variables that indicate
when, according to the founder,
sales were achieved for the first
time. The start-ups are assigned to
periods due to low case numbers for
single years (10 categories/dummy
variables for start-ups founded
between 1990 and 2007)
Industry Dummy variables indicating
industry the company is assigned to
(1-digit NACE level; 4 categories/
dummy variables).
Regional degree of Dummy variables indicating type of
agglomeration region with respect to the degree of
agglomeration (7 categories/
dummy variables, 1: highest; 7:
lowest).
East-West Move Dummy variables that indicate
whether the founder lived in East
Germany (West Germany) in 1989
and in West Germany (East
Germany) in 2010 (1: East to West;
2: West to East)
Share of R&D Share of employees (in%) that deal
employees primarily with R&D-tasks
Use of state-of-the-art Dummy variables that indicate
technology whether the technical equipment of
the company is state-of-the-art
(highest category) or almost state-of-
the-art (second-highest category).
Level of education The variable measures the highest
vocational training or university
degree founders had at the time of
their foundation (Scale: 1: lowest; 7
highest).
Industry experience Number of years the founder had
already worked in the branch in
which their company is active at the
time of the foundation.
Employment status Dummy variables indicating the
before start-up employment status. Status is
distinguished by 3 categories/
dummy variables (1; paid-
employment; 2: self-employment, 3:
other).
Self-employed role Dummy variable indicating whether
models at least one parent or close relatives
or friends had been self-employed
before starting the current business.
Number of (co-) Dummy variables that capture the
founders number of founders who
participated in the establishment of
the surveyed start-ups. This also
includes the interviewed founder (5
categories/dummy variables).
Research Policy 51 (2022) 104409
Table 1 (continued )
Name Description Source
Source
Takeover after start- Dummy variable that specifies GAW-ADIAB
GAW-ADIAB up whether the contacted company was
first established independently and
then taken over by another
company.
Lichter et al. Years in self- Number of years the founder spent GAW-ADIAB
(2021); Own employment before in self-employment before start-up
calculations start-up of current business.
Engaging in R&D Dummy variable that indicates GAW-ADIAB
activity whether the company of the founder
is engaged in research and
GAW-ADIAB development. Use in sample
selection stage only.
Openness Dummy variables indicating GAW-ADIAB
GAW-ADIAB whether founder is open for new
ideas and experimentation (founder
GAW-ADIAB with category 5–7 on a scale from
category 1–7). Use in sample
GAW-ADIAB selection stage only.
GAW-ADIAB
where the interviewed founder grew up.4 This is a proxy for the his­
torical effect of spying intensity on local trust levels at the time of the
entrepreneur’s initial socialization and imprinting. Since our proxy is
not measured at the individual level, but reflects a historical imprinting
effect long before respondents had the option to start a firm and to
cooperate on R&D, we circumvent capturing an endogenous effect of
GAW-ADIAB previous R&D cooperation and/or other business experiences feeding
back into founders’ level of trust.
Federal Statistical 2.3.3. Control variables
Office In the empirical analysis, we distinguish between founders who were
socialized in the former GDR from founders socialized in West Germany.
We introduce a dummy variable, GDR origin, indicating whether a per­
GAW-ADIAB son grew up in an East German region before 1989. Previous research
demonstrated that the socialist legacy left an imprint on East Germans’
mentality and preferences. For example, there is evidence that in­
dividuals exposed to socialism have a lower degree of trust in markets
and rely more heavily on the state (Corneo and Gruener, 2002; Alesina
GAW-ADIAB and Fuchs-Schuendeln, 2007; Bauernschuster et al., 2012). The GDR
origin dummy captures a variety of East-West differences and general
GAW-ADIAB
socialist legacy effects that may affect cooperative behavior. Hence, we
control for all these differences and isolate the general effect of being
exposed to a socialist regime on R&D cooperation from our main inde­
pendent variable (spying intensity) that measures regional variation in
GAW-ADIAB levels of surveillance and control to enforce ideology within the
authoritarian regime.
In our context, we have no clear expectations regarding the general
“GDR” effect. In the East German transition context, the socialist legacy
GAW-ADIAB is associated with lower levels of trust. At the same time, East Germans
lack experience with a market economy, and cooperation was not
common in socialist innovation systems (Mayntz, 1998). Decreased trust
GAW-ADIAB should reduce the willingness to engage in cooperative R&D, but lack of
experience increases the need to be cooperative. Thus, the net “GDR”
effect may depend on how pronounced the different underlying mech­
anisms are. We also introduce a dummy variable that identifies re­
spondents that were socialized in the GDR but reside in West Germany at
GAW-ADIAB the time of the interview, and a dummy variable marking West Germans
living in East Germany at the time of the interview (East-West Move).
We consider several other control variables that may affect the
GAW-ADIAB likelihood of being involved in R&D cooperation. At the individual level,
4
Please note that the region where the founder grew up can be a different
one from the region where the founder starts his or her venture. Hence, spying
intensity is not carrying effects on interfirm R&D cooperation specific to the
regions in which the start-up resides.
4
M. Wyrwich et al. Research Policy 51 (2022) 104409

Table 2 2.4. Empirical model

Summary statistics1.
Mean S.D. Min Max In our assessment, we deal with the issue of potential sample selec­
tion. For example, many service firms do not formally invest in R&D, but
Interfirm R&D cooperation (Dummy, Y = 1) 0.636 0.482 0 1
GDR origin (Dummy, Y = 1) 0.452 0.498 0 1 this does not mean that they do not engage in R&D collaborations. One
Spying intensity (in%) 0.172 0.21 0 0.742 key feature of services firms is that they co-create knowledge with their
West 0 0 0 0 clients without investing in R&D themselves. Therefore, R&D coopera­
East 0.381 0.132 0.166 0.742 tion among service firms is a non-observed outcome (i.e. no true zeros)
Age 47.465 6.433 37 61
Women (Dummy, Y = 1) 0.132 0.339 0 1
as they might be involved in a R&D cooperation without doing R&D
Start-up year 2001 4.433 1990 2007 themselves. In addition, there might be other factors that are either
Start-up size 2.11 2.462 1 20 controlled for or unobserved that may lead to selection into R&D ac­
Industry 5 1.967 2 7 tivity. Therefore, we can consider non-random selection of firms into
Regional degree of agglomeration 3 1.872 1 7
doing R&D. To this end, we present results from a two-stage Probit
Share of R&D employees (in%) 72.028 43.674 0 100
Use of state-of-the-art technology (Dummy, 0.632 0.483 0 1 model with sample selection.5 In the first stage, we use information on
Y = 1) the openness of the founder to experiment (openness, for a definition see
Level of Education 4.887 1.646 1 7 Table 1) to predict whether firms engage in R&D activities. This is
Industry experience 10.098 8.124 0 39 captured by a variable indicating the entrepreneur’s preference for new
Employment status before start-up: Other 0.097 0.296 0 1
(Dummy, Y = 1)
ideas and experiments. Based on the psychological Temperament and
Employment status before start-up: Paid 0.767 0.423 0 1 Character Inventory (TCI) that assesses personality, novelty seeking
employee (Dummy, Y = 1) (openness) and cooperation behavior are two distinct (non-related)
Employment status before start-up: Self- 0.137 0.344 0 1 personality domains (e.g., Garcia et al. 2017).6 The correlation between
employed (Dummy, Y = 1)
openness and R&D cooperation is also virtually zero (Table 3). Thus,
Self-employed role models before start-up 0.586 0.493 0 1
(Dummy, Y = 1) from a conceptual and empirical point of view, the exclusion restriction,
Takeover after start-up (Dummy, Y = 1) 0.155 0.362 0 1 namely that openness is not directly related to R&D cooperation, holds.
Number of founders 1.525 0.883 1 5 This is an important criterion for the correct use of sample selection
Years in self-employment before start-up 3 6.179 0 32 models (Certo et al., 2016). In Appendix B, we show that the results from
East-to-West Move (Dummy, Y = 1) 0.014 0.12 0 1
the selection stage indicate that our proxy for openness is a valid in­
West-to-East Move (Dummy, Y = 1) 0.093 0.291 0 1
Openness (Dummy, Y = 1) 0.447 0.498 0 1 strument. The analyses presented in Appendix B also suggest that sample
selection is not a severe issue.
Since it was also possible to be self-employed in the GDR, the number of years in
previous self-employment can also be rather large among East German founders.
3. Results

we consider the sociodemographic factors age and gender (women). We

also control for the level of education by the highest obtained vocational
degree (educational degree), and we consider industry experience in years.
We control for the employment status before start-up (a categorical vari­
able with the categories “paid employment,” “self-employment,” and
“other”), whether a founder had parents or other family members and
close friends who were self-employed before their own start-up (Self-
employed role models before start-up), the previous experience in self-
employment (years in self-employment before start-up), and the number
of (co-)founders. These individual factors can be important indicators of
venture performance (Bryant, 2014; Hahn et al., 2019; Grilli et al.,
2020), and may also influence strategic choices.
At the organizational level, we include the start-up period, start-up
size in terms of employment, and a variable that captures whether the
venture is already taken over by another venture after start-up (start-up
period, start-up size, and takeover after start-up). At the regional level, we
control for the degree of agglomeration, since firms in dense areas might
be more attractive cooperation partners due to (unobserved) differences
in productivity. We also control for industry effects at the 1-digit NACE
level. The decision to cooperate is not only based on trust, but is also
dependent on technology-related considerations. Such differences are
already partially captured by industry dummies. However, we also
control for the technical state of the machinery and equipment used by
the firm (use of state-of-the-art technology) and the share of employees
that are predominantly working in R&D (share of R&D employees), as this
might determine the likelihood of R&D cooperation as well.
Table 1 provides a definition of variables. In Table 2, we show
summary statistics for the sample of firms conducting R&D activities
(engaging in R&D activity, N = 214), and therefore facing the strategic
choice to cooperate or not. Table 3 presents the correlation matrix for
this sample. Summary statistics and correlation matrices for the full
sample that includes firms with no formal R&D activities (N = 620) are
provided in Appendix A (Tables A1 & A2).
3.1. Main results
Table 4 presents our main results. To rule out that our results are
driven by multicollinearity (Kalnins, 2018), we present models with
different sets of control variables to check for the robustness of our main
variable of interest. In model (1), we control for start-up size, gender,
age, and dummy variables for start-up year, industry, and region. We
also include the dummy variable to capture GDR origin but not the level
of spying intensity. The GDR origin dummy in model (1) is not signifi­
cant, revealing that there is no general negative GDR effect on interfirm
R&D cooperation. As explained in Section 2.3.3, this result does not
come as a surprise as opposing mechanisms may cancel each other out.
In model (2), we add the full set of individual controls. There is also no
significant coefficient estimate for the GDR origin dummy in this spec­
ification. The same applies to East-West-movers in model (1) and (2).
In Model 3, we introduce our measure for spying intensity along with
the GDR origin dummy. Hence, we control for the general GDR origin
effect on cooperation behavior and investigate the extent to which local
variation in trust-inhibiting spying activity affects interfirm R&D
cooperation. The results of Model 3 indicate that the level of spying
intensity experienced by founders has a negative effect on the likelihood
that their firms will engage in R&D cooperation. In terms of (average)
marginal effect size, an increase in the local share of the population
involved in spying intensity by 0.1 percentage points reduces the
5
The described context speaks in favor of a Heckman model instead of, for
example, a Tobit model (Amore and Murtinu, 2019, Fig. 2).
6
Novelty seeking is part of the habit system that is associated with temper­
ament which is linked to unconscious automatic reactions while cooperation is
a cognitive set related to character affecting conscious self-aware plans. At a
more abstract level, character and temperament are interrelated and form
identity. There is, however, no direct link between novelty seeking and coop­
eration in this model.

5
 Table 3

M. Wyrwich et al.
Correlation matrix.
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]

[1] Interfirm R&D 1

cooperation
[2] GDR origin 0.087 1
[3] Age 0.023 0.008 1
[4] Women − 0.117 0.075* 0.001 1
*
[5] Start-up period 0.078 − 0.032 − 0.188*** − 0.085** 1
[6] Start-up size 0.025 0.063 0.053 − 0.012 − 0.076* 1
[7] Industry − 0.092 − 0.139*** 0.052 0.06 − 0.069* − 0.223*** 1

[8] Degree of 0.101 0.319*** − 0.023 0.048 0.005 0.07* − 0.123*** 1

agglomeration
[9] Share of R&D − 0.009 − 0.015 − 0.028 0.095** − 0.064 − 0.081** 0.081** − 0.003 1
employees
[10] Use of state-of- 0.034 0.048 − 0.001 − 0.004 − 0.04 0.01 0.116* − 0.038 0.068 1
the-art
technology
[11] Level of 0.161** 0.035 0.066 − 0.028 0.036 0.03 0.242*** − 0.005 − 0.183*** − 0.048 1
Education
[12] Industry 0.061 0.042 0.4*** − 0.049 0.214*** 0.007 − 0.126*** − 0.008 − 0.018 0.022 − 0.047 1
experience
[13] Employment − 0.043 − 0.067* − 0.055 0.05 − 0.134*** − 0.081** 0.03 − 0.016 0.064 − 0.152** − 0.057 − 0.162*** 1
status before
start-up: Other
[14] Employment 0.003 0.028 − 0.016 0.013 − 0.009 0.042 − 0.072* − 0.036 − 0.026 0.101 0.002 0.056 − 0.593*** 1
status before
6

start-up: Paid
employee
[15] Employment 0.028 0.024 0.067* − 0.058 0.125*** 0.018 0.063 0.058 − 0.023 − 0.006 0.047 0.071* − 0.13*** − 0.722*** 1
status before
start-up: Self-
employed
[16] Self-employed 0.042 − 0.07* − 0.088** 0.038 0.11*** − 0.073* 0.064 0.022 − 0.051 − 0.003 0.042 − 0.031 − 0.035 0.039 − 0.017 1
role models
before start-up
[17] Takeover after 0.045 0.032 0.051 − 0.009 − 0.032 0.09** 0.021 0.028 0.058 0.076 0.051 0.067* − 0.004 − 0.027 0.037 − 0.057 1
start-up
[18] Number of 0.071 − 0.02 − 0.017 0 − 0.034 0.148*** − 0.028 0 − 0.076* 0.006 0.069* − 0.07* − 0.065 − 0.038 0.103 0 0.028 1
founders
[19] Years in self- 0.043 0.035 0.153*** − 0.083** 0.113*** 0.042 0.094** 0.04 − 0.06 − 0.078 0.031 0.163 − 0.054 − 0.414*** 0.556*** − 0.016 0.032 0.054 1
employment
before start-up
[20] East-to-West − 0.075 0.133*** 0.042 0.112*** − 0.018 − 0.055 0.029 − 0.088** 0.016 0.091 − 0.049 0.008 0.006 0.035 − 0.048 − 0.062 0.023 − 0.057 − 0.059 1
Move
[21] West-to-East 0.043 − 0.292*** − 0.009 − 0.011 − 0.042 0.049 0.078* 0.159*** 0.086** 0.006 0.052 − 0.051 0.007 − 0.045 0.049 0.023 0.077* 0.091** 0.05 − 0.039 1
Move

Research Policy 51 (2022) 104409

[22] Openness − 0.009 0.003 0.084** − 0.021 − 0.025 − 0.037 0.032 − 0.03 − 0.182*** 0.052 − 0.018 0.05 − 0.031 − 0.002 0.029 0.029 − 0.035 0.009 0.011 0.054 0.023 1
[23] Spying intensity − 0.187 . − 0.043 − 0.023 − 0.011 0.03 − 0.066 0.417*** − 0.049 − 0.259** − 0.065 − 0.053 0.063 − 0.031 − 0.01 0.063 − 0.078 − 0.089 0.007 − 0.16*** . 0.003
[in%]+ *

Notes:.
***
p<0.01.
**
p<0.05.
*
p<0.10. + Since spying intensity is zero for West German respondents, the correlation for spying intensity refers to East German respondents only.
M. Wyrwich et al.
likelihood that a venture established by a founder who grew up in this
region is involved in an R&D cooperation by almost 9 percent. The re­
sults are robust when removing the full set of control variables (Model
4).
The GDR origin dummy becomes positive and significant when the
measure for spying activity is introduced, but cannot be easily inter­
preted because the GDR origin dummy in Models 3 and 4 measures the
GDR effect of the hypothetical situation that spying intensity was zero in
the East German region where the founder grew up.7 Even though there
is no such case, it is possible to estimate marginal effects of GDR origin
for different levels of spying intensity. In any case, note that the prob­
ability of founders who grew up in the GDR to cooperate on R&D is
constantly decreasing with an increasing level of spying intensity
(Fig. 1).
3.2. Additional analyses
While Models 1–4 control for general exposure to socialism (via the
GDR dummy), one may argue that the length of exposure (i.e., the
number of years the founder lived in the GDR) may also affect cooper­
ation behavior negatively and drive our results. To further analyze this
possibility, we provide additional insights into the role of exposure
duration to socialism in our analytical framework in Model 5 (Table 1).
To this end, we interact the GDR origin dummy with age. In this stan­
dard dummy-continuous interaction, the constitutive age variable cap­
tures the age effect for West German founders, while the interaction
term captures the differential effect for East Germans whose age reveals
the duration of their general exposure to socialism. The interaction term
is insignificant. Hence, age does not affect cooperation behavior among
East Germans differently when compared to West Germans, while the
level of exposure to trust-inhibiting surveillance and control within the
regime is robust. Hence, the results from Model 5 indicate that our
findings are not driven by the length of time an individual was exposed
to socialism.8
In a further analysis, we extend the sample to include individuals
born after 1973. These founders were partly raised in post-unification
Germany in an environment where they were no longer exposed to
spying activity. When including these younger founders, we no longer
see an effect of spying intensity. This gives an indication that our main
results are not driven by region-specific effects, but by the exposure to a
high level of surveillance imposed by an authoritarian regime.9
4. Discussion and conclusions
Our study provides evidence that firm-level decisions regarding open
innovation are influenced by founders’ past experiences. More specif­
ically, our results reveal that firms founded by individuals who were
exposed to trust-reducing local contexts are less likely to engage in
interfirm R&D cooperation. We make use of a unique database to
identify regional differences in trust-inhibiting surveillance activities in
an authoritarian regime, and to explore how individual imprinting
associated with the experience of being exposed to such oppressive
regime enforcement practices influences the strategic decision making
7
This is the same interpretation as in a standard dummy-continuous inter­
action (for details, see Brambor et al. 2006).
8
Notwithstanding the arguments provided in Section 2.3.3, the level of
general exposure to socialism is an ambiguous measure for inhibiting trust,
since the GDR was relatively liberal, especially in the 1960s. The development
of an oppressive surveillance state and the application of strict socialist eco­
nomic principles did not occur until the later years of the GDR. For this reason,
an interaction between age and spying intensity would not be an appropriate
indicator in our analytical framework.
9
The results on the analyses, including younger entrepreneurs, can be ob­
tained upon request.
7
Research Policy 51 (2022) 104409
of founders.
Our study integrates insights from open innovation and imprinting
literature and contributes to these two literature streams in several
ways. First, we extend the understanding of how the micro-foundations
of open innovation (Bogers et al., 2018) help to explain firms’ propensity
to engage in open innovation strategies. We build on previous research
that highlights the effect of founders’ professional career backgrounds
(Hahn et al., 2019) on their firms’ engagement in open innovation
strategies, and show that the founder’s historical socioeconomic context
affects their strategic decision making long after the socioeconomic
context has changed.
Even though we do not measure trust directly, our findings based on
trust-inhibiting surveillance activities (Lichter et al., 2021) allow us to
join the discussion about the role of trust as a critical aspect of open
innovation and R&D collaboration (Carson et al., 2003; Brockman et al.,
2018). Innovation activities are characterized by endogeneity and un­
certainty. In collaborative settings, there is a risk of free-riding and value
appropriation conflicts created by disputes about how to allocate out­
puts (Larsson et al., 1998; Agarwal et al., 2010). Moreover, in the
context of open innovation, these challenges are difficult to overcome
using contracts (Mayer and Nickerson, 2005; Ulset, 1996). Hence,
collaborating in open innovation activities implies accepting some form
of vulnerability towards external partners. Going beyond existing
studies that often focus on relational trust between the partners to
develop theories about and assess the performance of collaborations (e.
g., Faems et al. 2008), we stress the importance of generalized trust
levels. This slightly different and complementary perspective expands
the discussion to the potential readiness of individuals to cooperate with
each other (Rothstein and Stolle, 2002).
Acknowledging previous research (Marquis and Qiao, 2020; Hahn
et al., 2019; O’Sullivan et al., 2021), we recognize the influence of a
founder’s past experiences on decision making. Going beyond previous
imprinting literature studies (Marquis and Tilcsik, 2013; Simsek et al.,
2015), we use a more fine-grained analysis of the founder’s socioeco­
nomic context to gain a deeper understanding of the imprinting process.
Specifically, we isolate and operationalize the regional differences of an
authoritarian regime’s specific enforcement mechanism (surveillance)
and interpret the imprinting effect of this mechanism on a firm’s stra­
tegic decision making. This approach separates the effect of oppressive
surveillance from other effects of being exposed to the authoritarian
regime, i.e., ideological imprinting. This more fine-grained analysis also
deepens our understanding of the processes that determine the
imprinting of certain characteristics on individuals and their organiza­
tions (Simsek et al., 2015).
Our analyses show that research on authoritarian imprinting must
disentangle ideological imprinting and imprinting stemming from
oppressive regime enforcement practices. Ideological imprinting refers
to the process of imprinting a “perceptual framework of interrelated
ideas and beliefs” (Wang et al., 2019, p. 1113). Although Marquis and
Qiao (2020) document the important role ideological imprinting plays
in strategic decision-making, our study shows that the oppressive ac­
tions of authoritarian regimes also leave an imprint that is different than
(and independent of) ideological imprinting. Accounting for both types
of imprinting and taking a fine-grained regional look at the socioeco­
nomic factors leading to individual imprinting is clearly warranted in
future research.
Despite the specific nature of our empirical setting, we are confident
that our conclusions can inform discussions in very different contexts.
Based on the notion that generalized trust levels are mostly sticky over
time (Algan and Cahuc, 2010), we assume that our insights are relevant
in other contexts and can deepen our understanding of how similar
transitions from an oppressive authoritarian regime to more open soci­
eties in other countries impacted their citizens (e.g., Chile, Argentina,
Spain, Czech Republic, etc.).
It is important for policy-makers and regional developers to recog­
nize that regional and institutional contexts leave a heterogenous
M. Wyrwich et al. Research Policy 51 (2022) 104409

Table 4
The relationship between East German origin and R&D cooperation behavior.
1 2 3 4 5

Main results Robustness checks

Spying intensity (in%) − 0.867** − 0.894** − 0.887**

(0.361) (0.366) (0.392)
GDR origin (Dummy, Y = 1) 0.032 0.051 0.370** 0.358** − 0.147
(0.093) (0.093) (0.168) (0.168) (0.403)
Age 0.005 0.005 0.006 0.006 0.001
(0.005) (0.005) (0.005) (0.005) (0.006)
GDR origin (Dummy, Y = 1) x Age 0.011
(0.009)
Women (Dummy, Y = 1) − 0.040 − 0.026 0.001 − 0.002 0.011
(0.145) (0.141) (0.124) (0.121) (0.117)
Start-up size 0.001 − 0.001 0.003 0.003 0.001
(0.013) (0.012) (0.011) (0.012) (0.011)
Start-up period (n = 10) (Dummies) Yes Yes Yes Yes Yes
Industry (1-digit) (n = 4) (Dummies) Yes Yes Yes Yes Yes
Regional degree of agglomeration (n = 6) (Dummies) Yes Yes Yes Yes Yes
East-to-West Move (Dummy, Y = 1) − 0.261 − 0.246 − 0.350* − 0.349 − 0.365*
(0.244) (0.231) (0.206) (0.213) (0.198)
West-to-East Move (Dummy, Y = 1) 0.027 0.016 0.006 0.020 0.003
(0.128) (0.138) (0.129) (0.117) (0.123)
Share of R&D employees (in%) − 0.001 − 0.001 − 0.001
(0.001) (0.001) (0.001)
Use of state-of-the-art technology (Dummy, Y = 1) 0.012 − 0.004 − 0.008
(0.058) (0.057) (0.054)
Level of Education 0.014 0.004 − 0.005
(0.035) (0.030) (0.028)
Industry experience − 0.000 − 0.001 − 0.001
(0.004) (0.004) (0.004)
Employment status before start-up: Other Yes Yes Yes
(Reference group)
Employment status before start-up: Employee − 0.011 − 0.022 − 0.018
(0.102) (0.098) (0.093)
Employment status before start-up: Self-employed 0.014 0.005 0.013
(0.125) (0.120) (0.111)
Self-employed role models before start-up 0.043 0.042 0.043
(Dummy, Y = 1) (0.067) (0.064) (0.063)
Takeover after start-up (Dummy, Y = 1) 0.095 0.092 0.077
(0.082) (0.080) (0.077)
Years in self-employment before start-up − 0.003 − 0.003 − 0.003
(0.005) (0.005) (0.005)
Number of founders (n = 5) (Dummies) No Yes Yes No Yes
Wald Chi2 192.80*** 249.96*** 195.70*** 128.77*** 114.96***
Observations (Total) 620 618 620 620 618
Observations (Selected) 214 212 212 214 212

Notes: Probit model with sample selection. First stage results are provided in Table B1. Robust standard errors in parentheses.
***
p<0.01.
**
p<0.05.
*
p<0.10. Average marginal effects reported (unconditioned on selection). Constant not reported for brevity. There are two missing values on the R&D employment
share. This explains the minor difference in case numbers across models.

imprint on founders that influence their firms’ strategic choices. Because
barriers to R&D cooperation and the relative economic success of firms
and regions may be rooted in history and depend on the intensity of
specific historical events, policy-makers need to address “soft” barriers
to R&D cooperation to promote firm- and regional-level development.
Our findings suggest that founders from authoritarian regimes who do
not transition well to a market economy may impose imprinted biases on
their ventures and ignore the full range of strategic options free-market
economies offer. Policies that support awareness campaigns and trust-
facilitating networking might help founders be more aware of and
address personal biases that negatively affect firm-level decision-
making.
Our research has limitations that future research could address. We
were not able to explicitly consider the detailed composition of the
founding team (e.g., Bryant 2014, Hahn et al. 2019). Future research
might look at the heterogeneity of founders in order to investigate how
the combination of different imprints influences strategic choices in new
ventures. Our study does not investigate how a founder’s imprint is
transferred to the organization, nor whether a founder’s imprint affects
the strategic decisions of the firm after the founder leaves the organi­
zation. Because our study’s sample includes mostly smaller firms, it is
difficult to generalize our findings to large firms. In large firms, strategic
decisions are usually made by management teams, and in that context,
relational trust may be more important than generalized trust.
An interesting avenue for future research is investigating the inter­
action between ideological imprinting (cf., Marquis and Qiao 2020) and
oppressive surveillance in empirical contexts where certain regions,
organizations or people were differently exposed to the ideology of the
authoritarian regime. It is also interesting to explore the interaction
between pre-existing local cultures and exposure to authoritarian re­
gimes (for a general discussion in the context of Germany, see Becker
et al. 2020). Our findings might also encourage future research on how

8
M. Wyrwich et al. Research Policy 51 (2022) 104409

Fig. 1. The marginal effect of spying intensity on cooperation behavior (with 95% confidence interval, CI).

the socioeconomic context of immigrant founders (cf. Kerr and Kerr Appendix A
2020) influences their willingness to participate in open innovation
activities. Table A1:
Our data has only limited information about the scope of R&D ac­ Table A2:
tivities and the technology-related contexts in which firms operate, an
issue that future research might address. Additional research into the
role of R&D contracts that might mitigate trust-related obstacles for Table A1
cooperation is also called for. Future research may also want to explore Summary statistics: Full sample.
how exogenous measures affecting local trust levels link to endogenous
Mean S.D. Min Max
trust formation at the individual level. Here, a better understanding of
the interrelation between socioeconomic imprinting and the emergence Engaging in R&D activity (Dummy, Y = 1) 0.351 0.478 0 1
GDR origin (Dummy, Y = 1) 0.491 0.5 0 1
of generalized trust and relational trust could be an especially promising
Spying intensity (in%) 0.185 0.209 0 0.742
avenue for future research in the context of studying collaborative West 0 0 0 0
behavior. East 0.377 0.127 0.166 0.742
Age 45.582 7.763 25 61
Women (Dummy, Y = 1) 0.144 0.351 0 1
Funding
Start-up year 2001 4.314 1990 2007
Start-up size 2.099 2.41 1 20
This research did not receive any specific grant from funding Industry 5 1.941 2 7
agencies in the public, commercial, or not-for-profit sectors. Regional degree of agglomeration 3 1.833 1 7
Level of Education 4.824 1.693 1 7
Industry experience 9.27 8.011 0 39
CRediT authorship contribution statement
Employment status before start-up: Other 0.127 0.334 0 1
(Dummy, Y = 1)
Michael Wyrwich: Conceptualization, Methodology, Writing – Employment status before start-up: Paid 0.736 0.441 0 1
original draft, Writing – review & editing, Formal analysis. Philip J. employee (Dummy, Y = 1)
Steinberg: Conceptualization, Methodology, Writing – original draft, Employment status before start-up: Self- 0.137 0.344 0 1
employed (Dummy, Y = 1)
Writing – review & editing. Florian Noseleit: Conceptualization, Self-employed role models before start-up 0.601 0.49 0 1
Methodology, Writing – original draft, Writing – review & editing. (Dummy, Y = 1)
Pedro de Faria: Conceptualization, Methodology, Writing – original Takeover after start-up (Dummy, Y = 1) 0.145 0.353 0 1
draft, Writing – review & editing. Number of founders 1.55 0.894 1 5
Years in self-employment before start-up 2.807 5.905 0 32
East-to-West Move (Dummy, Y = 1) 0.014 0.118 0 1
Declaration of Competing Interest West-to-East Move (Dummy, Y = 1) 0.085 0.28 0 1
Openness (Dummy, Y = 1) 0.443 0.497 0 1
None.
Notes: Interfirm R&D cooperation, share of R&D-employees, and use of state-of-
the-art-technology was only asked if the company was involved in R&D
activities.

9
 M. Wyrwich et al.
Table A2
Correlation Matrix: Full sample.
[1] [2] [3] [4] [5] [6] [7] [8] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

[1] R&D Activity 1

[2] GDR origin − 0.001 1
[3] Age − 0.02 − 0.118*** 1
[4] Women − 0.11*** 0.123*** − 0.051 1
[5] Start-up period 0.049 0.009 − 0.239*** − 0.046 1
[6] Start-up size 0.068* 0.051 0.047 − 0.017 − 0.073* 1
[7] Industry − 0.089** − 0.116*** 0.04 0.056 − 0.07* − 0.228*** 1
[8] Degree of agglomeration − 0.024 0.312*** − 0.025 0.072* 0.017 0.07* − 0.122*** 1
[9] Level of Education 0.151*** 0.021 0.122*** 0.019 0.02 0.041 0.219*** 0.005 1
[10] Industry experience 0.004 − 0.011 0.464*** − 0.069* 0.159*** 0.01 − 0.115*** − 0.017 − 0.001 1
[11] Employment status before start-up: Other − 0.043 − 0.014 − 0.19*** 0.046 − 0.057 − 0.077** 0.001 0.012 − 0.102*** − 0.223*** 1
[12] Employment status before start-up: Paid 0.002 − 0.008 0.111*** 0.011 − 0.037 0.037 − 0.04 − 0.041 0.044 0.126*** − 0.637*** 1
employee
[13] Employment status before start-up: Self- 0.039 0.024 0.042 − 0.059 0.103*** 0.027 0.05 0.041 0.042 0.054 − 0.152*** − 0.665*** 1
employed
10

[14] Self-employed role models before start-up 0.042 − 0.046 − 0.123*** 0.041 0.118*** − 0.055 0.048 0.024 0.038 − 0.061 0.011 − 0.009 0.001 1
[15] Takeover after start-up − 0.029 0.023 0.084** − 0.011 − 0.043 0.082** 0.011 0.035 0.052 0.072* − 0.027 − 0.005 0.032 − 0.053 1
[16] Number of founders 0.131*** − 0.019 − 0.056 − 0.012 − 0.023 0.134*** − 0.025 − 0.017 0.065* − 0.081** − 0.028 − 0.071* 0.119*** 0.002 0.013 1
[17] Years in self-employment before start-up 0.073* 0.006 0.166*** − 0.097*** 0.092** 0.043 0.09** 0.031 0.04 0.172*** − 0.064* − 0.372*** 0.54*** − 0.011 0.034 0.049 1
[18] East-to-West Move − 0.013 0.121*** 0.037 0.121*** − 0.016 − 0.054 0.034 − 0.062* − 0.058 0.005 − 0.01 0.017 − 0.013 − 0.049 0.018 − 0.047 − 0.055 1
[19] West-to-East Move − 0.057 − 0.3*** 0.039 − 0.025 − 0.048 0.05 0.064* 0.152*** 0.055 − 0.031 − 0.011 − 0.021 0.038 0.024 0.073* 0.087** 0.051 − 0.036
[20] Openness 0.218*** − 0.019 0.068* − 0.041 − 0.033 − 0.033 0.045 − 0.053 − 0.008 0.055 − 0.053 0.011 0.038 0.044 − 0.048 0.022 0.023 0.038
[21] Spying intensity [in%]+ 0.035 0.011 − 0.027 − 0.013 0.04 − 0.089* 0.404*** − 0.073 − 0.017 0.009 0.009 − 0.02 0.057 − 0.07 − 0.129** 0.007 − 0.118** .

Notes:.
***
p<0.01.
**
p<0.05.
*
p<0.10. + Since spying intensity is zero for West German respondents, the correlation for spying intensity refers to East German respondents only. Interfirm R&D cooperation, share of R&D-employees, and use of state-
of-the-art-technology was only asked if the company was involved in R&D activities.

Research Policy 51 (2022) 104409

M. Wyrwich et al. Research Policy 51 (2022) 104409

Appendix B
The models in Table B1 show that openness to new ideas and
experimentation is positively linked to the likelihood of engaging in
R&D. Hence, the data suggests that this is a valid instrument in the two-
stage model with our sample selection. Some of the control variables,
such as gender and education, also determine selection into R&D ac­
tivities. Nevertheless, sample selection does not seem to be a severe issue
as the correlation coefficient (rho) between the error term of the selec­
tion model and the second-stage model is relatively low and insignifi­
cant, which is indicated by the values for athrho, the inverse hyperbolic
tangent of rho. Interestingly, spying intensity and the GDR origin
dummy are both not significantly related to engaging in R&D activities.
The models presented in Table 4 do not include an inverse mills ratio.
This is because of the binary nature of the dependent variable (interfirm
R&D cooperation). In a Probit model with sample selection the entire
model is estimated via maximum likelihood without inverse mills ratio
correction. This is in contrast to a standard Heckman model where the
dependent variable is continuous. As a consistency check, we ran a
standard two-step Heckman model and calculate the inverse mills ratio
despite the fact that our dependent variable is not continuous. The re­
sults are provided in Table B2. The results for our main variables of
interest remain robust while the inverse mills ratio is insignificant,
implying that sample selection is not a severe issue.

Table B1
Selection stage results: Probit with sample selection.
Dep var: Engaging in R&D activities (Dummy, Y = 1) 1 2 3 4 5

Openness 0.174*** 0.175*** 0.172*** 0.172*** 0.172***

(0.035) (0.034) (0.034) (0.035) (0.034)
Spying intensity (in%) 0.349 0.223 0.351
(0.232) (0.234) (0.232)
GDR origin (Dummy, Y = 1) − 0.027 − 0.062 − 0.188* − 0.107 − 0.199
(0.066) (0.065) (0.104) (0.107) (0.289)
Age 0.001 0.001 0.001 0.001 0.001
(0.003) (0.003) (0.003) (0.003) (0.004)
GDR origin (Dummy, Y = 1) x Age 0.000
(0.006)
Women (Dummy, Y = 1) − 0.153*** − 0.133** − 0.133*** − 0.153*** − 0.133**
(0.051) (0.052) (0.052) (0.050) (0.052)
Start-up size 0.015** 0.009 0.009 0.015** 0.009
(0.008) (0.008) (0.008) (0.008) (0.008)
Start-up period (n = 10) (Dummies) Yes Yes Yes Yes Yes
Industry (1-digit) (n = 4) (Dummies) Yes Yes Yes Yes Yes
Regional degree of agglomeration (n = 6) (Dummies) Yes Yes Yes Yes Yes
East-to-West Move (Dummy, Y = 1) 0.037 0.138 0.165 0.055 0.166
(0.158) (0.154) (0.158) (0.163) (0.158)
West-to-East Move (Dummy, Y = 1) − 0.138 − 0.187** − 0.175** − 0.129 − 0.176**
(0.086) (0.087) (0.086) (0.086) (0.087)
Level of Education 0.061*** 0.062*** 0.061***
(0.013) (0.012) (0.013)
Industry experience − 0.003 − 0.003 − 0.003
(0.003) (0.003) (0.003)
Employment status before start-up: Other Yes Yes Yes
(Reference group)
Employment status before start-up: Employee 0.023 0.026 0.027
(0.063) (0.062) (0.062)
Employment status before start-up: Self-employed − 0.014 − 0.011 − 0.010
(0.081) (0.080) (0.080)
Self-employed role models before start-up 0.020 0.017 0.017
(Dummy, Y = 1) (0.037) (0.037) (0.037)
Takeover after start-up (Dummy, Y = 1) − 0.012 − 0.007 − 0.007
(0.050) (0.050) (0.050)
Years in self-employment before start-up 0.005 0.005 0.005
(0.003) (0.003) (0.003)
Number of founders (n = 5) (Dummies) Yes Yes Yes
Atrho − 0.267 − 0.247 − 0.296 − 0.355 − 0.401
(0.547) (0.610) (0.553) (0.514) (0.636)
Observations 620 618 620 620 618

Notes: First stage results from Probit model with sample selection. Robust standard errors in parentheses.
***
p<0.01.
**
p<0.05.
*
p<0.10. Average marginal effects reported. Constant not reported for brevity. There are two missing values on the R&D employment share. This explains the minor
difference in case numbers across models. Atrho measures the correlation between error terms in the first and second stage of the selection model. Based on a Wald test,
the null hypothesis that this correlation is zero cannot be rejected across model specifications. Hence, selection is likely not problematic.

11
M. Wyrwich et al. Research Policy 51 (2022) 104409

Table B2
Second stage results: Standard Heckman twostep model.
Dep Var: Interfirm R&D Cooperation (Dummy, Y = 1) 1 2 3 4 5

Spying intensity (in%) − 0.853** − 0.948*** − 0.901**

(0.392) (0.362) (0.389)
GDR origin (Dummy, Y = 1) 0.030 0.038 0.349* 0.374** − 0.207
(0.112) (0.122) (0.190) (0.172) (0.568)
Age 0.005 0.004 0.005 0.005 − 0.000
(0.005) (0.006) (0.006) (0.005) (0.008)
GDR origin (Dummy, Y = 1) x Age 0.012
(0.011)
Women (Dummy, Y = 1) − 0.062 − 0.037 0.001 − 0.012 0.006
(0.167) (0.174) (0.174) (0.167) (0.175)
Start-up size 0.001 − 0.001 0.001 0.003 − 0.001
(0.014) (0.014) (0.015) (0.014) (0.015)
Inverse mills ratio − 0.087 − 0.100 − 0.136 − 0.129 − 0.153
(0.188) (0.194) (0.197) (0.191) (0.198)
Start-up period (n = 10) (Dummies) Yes Yes Yes Yes Yes
Industry (1-digit) (n = 4) (Dummies) Yes Yes Yes Yes Yes
Regional degree of agglomeration (n = 6) (Dummies) Yes Yes Yes Yes Yes
Controls Table 4 column II Yes Yes Yes
F-Statistics 9.41*** 11.97*** 9.70*** 9.65*** 10.99***
Observations (Total) 620 618 620 620 618
Observations (Selected) 214 212 212 214 212

Notes: Standard two-stage heckman model. Robust standard errors in parentheses.

***
p<0.01.
**
p<0.05.
*
p<0.10. Average marginal effects reported (unconditioned on selection). Constant not reported for brevity. There are two missing values on the R&D employment
share. This explains the minor difference in case numbers across models.

References Chesbrough, H.W., 2006. Open Business Models: How To Thrive In The New Innovation
Landscape. Harvard Business Press, Boston.
Chesbrough, H.W., Bogers, M., 2014. Explicating Open Innovation: clarifying an
Agarwal, R., Croson, R., Mahoney, J.T., 2010. The role of incentives and communication
Emerging Paradigm for Understanding Innovation. In: Chesbrough, H.,
in strategic alliances: an experimental investigation. Strateg. Manag. J. 31 (4),
Vanhaverbeke, W., West, J. (Eds.), New Frontiers in Open Innovation. Oxford
413–437.
University Press, pp. 3–28.
Alesina, A., Fuchs-Schuendeln, N., 2007. Good bye Lenin (or not?): the effect of
Childs, D., Popplewell, R., 1996. The Stasi: The East German Intelligence and Security
communism on people’s preferences. Am. Econ. Rev. 97, 1507–1528.
Service. New York University Press, New York, NY.
Algan, Y., Cahuc, P., 2010. Inherited trust and growth. Am. Econ. Rev. 100 (5),
CIS (2020). Community Innovation Survey, https://ec.europa.eu/eurostat/web/mi
2060–2092.
crodata/community-innovation-survey.
Amore, M.D., Murtinu, S., 2019. Tobit models in strategy research: critical issues and
Classen, N., Van Gils, A., Bammens, Y., Carree, M., 2012. Accessing resources from
applications. Glob. Strategy J. Online 1–25.
innovation partners: the search breadth of family SMEs. J. Small Bus. Manag. 50,
Aslund, A., Djankov, S., 2014. The Great Rebirth: Lessons from the Victory of Capitalism
191–215.
over Communism. The Great Rebirth: Lessons from the Victory of Capitalism over
Corneo, G., Gruner, H.P., 2002. Individual preferences for political redistribution.
Communism. Peterson Institute for International Economics.
J. Public Econ. 83 (1), 83–107.
Audretsch, D.B., Moog, P., 2020. Democracy and entrepreneurship. Entrep. Theory
Dahlander, L., Gann, D.M., 2010. How open is innovation? Res. Policy 39, 699–709.
Pract., 1042258720943307
De Faria, P., Lima, F., Santos, R., 2010. Cooperation in innovation activities: the
Bauernschuster, S., Falck, O., Gold, R., Heblich, S., 2012. The shadows of the socialist
importance of partners. Res. Policy 39, 1082–1092.
past: lack of self-reliance hinders entrepreneurship. Eur. J. Polit. Econ. 28 (4),
De Faria, P., Noseleit, F., Los, B., 2020. The influence of internal barriers on open
485–497.
innovation. Ind. Innov. 27, 205–209.
Becker, S.O., Mergele, L., Woessmann, L., 2020. The separation and reunification of
Delhey, J., Newton, K., 2005. Predicting cross-national levels of social trust: global
Germany: rethinking a natural experiment interpretation of the enduring effects of
pattern or nordic exceptionalism? Eur. Sociol. Rev. 21 (4), 311–327.
communism. J. Econ. Perspect. 34 (2), 143–171.
Dinesen, P.T., Bekkers, R., 2017. The foundations of individuals’ generalized social trust:
Bjørnskov, C., 2006. The multiple facets of social capital. Eur. J. Polit. Econ. 22 (1),
a review. In: Van Lange, P.A.M., Rockenbach, B., Yamagishi, T. (Eds.), Trust in Social
22–40.
Dilemmas. Oxford University Press, New York, NY, pp. 77–100.
Bogers, M., Foss, N.J., Lyngsie, J., 2018. The “human side” of open innovation: the role of
Ebersberger, B., Galia, F., Laursen, K., Salter, A., 2021. Inbound open innovation and
employee diversity in firm-level openness. Res. Policy 47 (1), 218–231.
innovation performance: a robustness study. Res. Policy 50, 104271.
Brambor, T., Clark, W.R., Golder, M., 2006. Understanding interaction models:
Eftekhari, N., Bogers, M., 2015. Open for entrepreneurship: how open innovation can
improving empirical analyses. Polit. Anal. 14, 63–82.
foster new venture creation. Creat. Innov. Manag. 24, 574–584.
Brockman, P., Khurana, I.K., Zhong, R.I., 2018. Societal trust and open innovation. Res.
Ertug, G., Cuypers, I.R., Noorderhaven, N.G., Bensaou, B.M., 2013. Trust between
Policy 47 (10), 2048–2065.
international joint venture partners: effects of home countries. J. Int. Bus. Stud. 44
Bryant, P.T., 2014. Imprinting by design: the microfoundations of entrepreneurial
(3), 263–282.
adaptation. Entrep. Theory Pract. 38 (5), 1081–1102.
Eurostat, 2018. Glossary: high-tech, https://ec.europa.eu/eurostat/statistics-explain
Bublitz, E., Nielsen, K., Noseleit, F., Timmermans, B., 2018. Entrepreneurship, Human
ed/index.php?title=Glossary:High-tech.
Capital, and Labor Demand: a Story of Signaling and Matching. Ind. Corp. Change
Faems, D., Janssens, M., Madhok, A., Looy, B.V., 2008. Toward an integrative
27, 269–287.
perspective on alli-ance governance: connecting contract design, trust dynamics, and
Burns, N., Kinder, D., Rahn, W., 2003. Social trust and democratic politics. In:
contract application. Acad. Manag. J. 51 (6), 1053–1078.
Proceedings of the Paper prepared for the Annual Meeting of the Mid-West Political
Ferrin, D.L., Gillespie, N., 2010. Trust differences across national-societal cultures: much
Science Association. Chicago Illinois. Mid-West Political Science Association,
to do, or much ado about nothing? In: Saunders, M.N.K., Skinner, D., Dietz, G.,
Chicago. April 2003.
Gillespie, N., Lewicki, R.J. (Eds.), Organizational trust: A cultural Perspective.
Burton, M.D., Beckman, C.M., 2007. Leaving a legacy: position imprints and successor
Cambridge University Press, Cambridge, UK, pp. 42–86.
turnover in young firms. Am. Sociol. Rev. 72 (2), 239–266.
Ferriani, S., Lazerson, M.H., Lorenzoni, G., 2020. Anchor entrepreneurship and industry
Carson, S.J., Madhok, A., Varman, R., John, G., 2003. Information processing moderators
catalysis: the rise of the Italian Biomedical Valley. Res. Policy 49 (8), 104045.
of the effectiveness of trust-based governance in interfirm R&D collaboration. Organ.
Fuchs-Schuendeln, N., Masella, P., 2016. Long-lasting effects of socialist education. Rev.
Sci. 14 (1), 45–56.
Econ. Stat. 98, 428–441.
Certo, S.T., Busenbark, J.R., Woo, H.S., Semadeni, M., 2016. Sample Selection Bias and
Fritsch, M., Bublitz, E., Sorgner, A., Wyrwich, M., 2014. How much of a socialist legacy?
Heckman models in strategic management research. Strateg. Manag. J. 37,
The Re-emergence of entrepreneurship in the East German transformation to a
2639–2657.
market economy. Small Bus. Econ. 43, 427–446.
Chesbrough, H.W., 2003. Open Innovation: The New Imperative for Creating and
Profiting from Technology. Harvard Business Press, Boston.

12
M. Wyrwich et al.
Garcia, D., Lester, N., Cloninger, K.M., Cloninger, R.C., 2017. Temperament and
Character Inventory (TCI). In: Zeigler-Hill, V., Shackelford, T. (Eds.), Encyclopedia
of Personality and Individual Differences. Springer, Cham.
GEM, 2020. German National Report. (downloaded on 11 October 2021 from https
://gemconsortium.org/news/key-findings-from-the-2019%2F2020-german-nation
al-report).
Gerschewski, J., 2013. The three pillars of stability: legitimation, repression, and co-
optation in autocratic regimes. Democratization 20 (1), 13–38.
Giuliano, P., Spilimbergo, A., 2014. Growing up in a recession. Rev. Econ. Stud. 81,
787–817.
Glanville, J.L., Paxton, P., 2007. How do we learn to trust? A confirmatory tetrad
analysis of the sources of generalized trust. Soc. Psychol. Q. 70 (3), 230–242.
Grilli, L., Jensen, P.H., Murtinu, S., Park, H.D., 2020. A close look at the contingencies of
founders’ effect on venture performance. Ind. Corp. Change 29 (4), 997–1020.
Hahn, D., Minola, T., Eddleston, K.A., 2019. How do scientists contribute to the
performance of innovative start-ups? An imprinting perspective on open innovation.
J. Manag. Stud. 56, 895–928.
Hardin, R., 2006. Trust. Trust. Wiley, Cambridge, UK.
Huff, L., Kelley, L., 2003. Levels of organizational trust in individualist versus collectivist
societies: a seven-nation study. Organ. Sci. 14 (1), 81–90.
Kalnins, A., 2018. Multicollinearity: how common factors cause Type 1 errors in
multivariate regression. Strateg. Manag. J. 39, 2362–2385.
Kerr, S.P., Kerr, W., 2020. Immigrant entrepreneurship in America: evidence from the
survey of business owners 2007 & 2012. Res Policy 49 (3), 103918.
Keupp, M.M., Gassmann, O., 2009. Determinants and archetype users of open
innovation. R&D Manag. 39, 331–341.
Knabe, H., 2014. The dark secret of a surveillance state. In: Proceedings of the TED
Conferences (downloaded on 12 January 2021 from. https://www.ted.com/talks/hu
bertus_knabe_the_dark_secrets_of_a_surveillance_state/transcript.
Kramer, R.M., Lewicki, R.J., 2010. Repairing and enhancing trust: approaches to
reducing organizational trust deficits. Acad. Manag. Ann. 4, 245–277.
Larsson, R., Bengtsson, L., Henriksson, K., Sparks, J., 1998. The interorganizational
learning dilemma: collective knowledge development in strategic alliances. Organ.
sci. 9 (3), 285–305.
Krishnan, R., Martin, X., Noorderhaven, N.G., 2006. When does trust matter to alliance
performance? Acad. Manag. J. 49 (5), 894–917.
Lee, Y., Cavusgil, S.T., 2006. Enhancing alliance performance: the effects of contractual-
based versus relational-based governance. J. Bus. Res. 59 (8), 896–905.
Lee, S., Park, G., Yoon, B., Park, J., 2010. Open innovation in SMEs – An intermediated
network model. Res. Policy 39, 290–300.
Lichter, A., Loeffler, M., Siegloch, S., 2021. The Long-Term Costs of Government
Surveillance: insights from Stasi Spying in East Germany. J. Eur. Econ. Assoc. 19,
741–789.
Marquis, C., Qiao, K., 2020. Waking from Mao’s dream: communist ideological
imprinting and the internationalization of entrepreneurial ventures in China. Adm.
Sci. Q. 65 (3), 795–830.
Marquis, C., Tilcsik, A., 2013. Imprinting: Toward a multilevel Theory, 7. Academy of
Management Annals, pp. 195–245.
Mayer, R.C., Davis, J.H., Schoorman, D., 1995. An Integrative Model of Organizational
Trust. Acad. Manag. Rev. 20, 709–734.
13
Research Policy 51 (2022) 104409
Mayer, K.J., Nickerson, J.A., 2005. Antecedents and performance implications of
contracting for knowledge workers: evidence from information technology services.
Organ. Sci. 16 (3), 225–242.
Mayntz, R., 1998. Socialist academies of sciences: the enforced orientation of basic
research at user needs. Res. Policy 27 (8), 781–791.
Nannestad, P., 2008. What Have We Learned About Generalized Trust, If Anything?
Annu. Rev. Polit. Sci. 11, 413–436.
Nooteboom, B., Berger, H., Noorderhaven, N.G., 1997. Effects of trust and governance on
relational risk. Acad. Manag. J. 40 (2), 308–338.
O’Sullivan, D., Zolotoy, L., Fan, Q., 2021. CEO early-life disaster experience and
corporate social performance. Strateg. Manag. J. 42, 2137–2161 forthcoming.
Paldam, M., Svendsen, G.T., 2001. Missing social capital and the transition in Eastern
Europe. J. Inst. Innov. Dev. Trans. 5, 21–34.
Rainer, H., Siedler, T., 2009. Does democracy foster trust? J. Comp. Econ. 37 (2),
251–269.
Roche, M.P., Conti, A., Rothaermel, F.T., 2020. Different founders, different venture
outcomes: a comparative analysis of academic and non-academic startups. Res.
Policy 49 (10), 104062.
Rothstein, B., Stolle, D., 2002. How political institutions create and destroy social capital:
an institutional theory of generalized trust. In: Proceedings of the 98th meeting of
the American Political Science Association in. Boston, MA.
Rotter, J.B., 1971. Generalized expectancies for interpersonal trust. Am. Psychol. 26 (5),
443.
Simsek, Z., Fox, B.C., Heavey, C., 2015. What’s past is prologue” A framework, review,
and future directions for organizational research on imprinting. J. Manag. 41 (1),
288–317.
Srinivasan, R., Choo, A., Narayanan, S., Sarkar, S., Tenhiälä, A., 2021. Knowledge
sources, innovation objectives, and their impact on innovation performance: quasi-
replication of Leiponen and Helfat, 2010. Strateg. Manag. J. 42, 2104–2136
forthcoming.
Stinchcombe, A.L., 1965. Social structure and organizations. Handb. Organ. 7, 142–193.
Svolik, M.W., 2012. The Politics of Authoritarian Rule. Cambridge University Press, New
York, NY.
Ulset, S., 1996. R&D outsourcing and contractual governance: an empirical study of
commercial R&D projects. J. Econ. Behav. Organ. 30 (1), 63–82.
Tabellini, G., 2010. Culture and institutions: economic development in the regions of
Europe. J. Eur. Econ. Assoc. 8 (4), 677–716.
Waguespack, D.M., Dunford, E.T., Birnir, J.K., 2018. Cultural imprinting, institutions,
and the organization of new firms. Strategy Sci. 3 (2), 426–438.
Wang, D., Du, F., Marquis, C., 2019. Defending Mao’s dream: how politicians‘ ideological
imprinting affects firms‘ political appointment in China. Acad. Manag. J. 62,
1111–1136.
West, J., Salter, A., Vanhaverbeke, W., Chesbrough, H., 2014. Open innovation: the next
decade. Res. Policy 43, 805–811.
Wyrwich, M., Fritsch, M., Bublitz, E., Sorgner, A., 2018. GAW survey data linked with
administrative data of the IAB (GAW-ADIAB): a documentation and review of
empirical evidence and research opportunities. JENA Econ. Res. Paper #2018-017.
Zaheer, A., Mcevily, B., Perrone, V., 1998. Does Trust Matter? Exploring the Effects of
Interorganizational and Interpersonal Trust on Performance. Strateg. Manag. J. 9
(2), 141–159.