Journal of Economic Behavior & Organization 95 (2013) 90–110
Contents lists available at ScienceDirect
Journal of Economic Behavior & Organization
journal homepage: www.elsevier.com/locate/jebo
Institutions, culture, and open source夽
Sebastian v. Engelhardt a,∗ , Andreas Freytag a,b
a
b
Friedrich-Schiller-University Jena, Department of Economics and Business Administration, Carl-Zeiss-Str. 3, 07743 Jena, Germany
Stellenbosch University, Department of Economics, Private Bag x1 Matieland, 7602 Stellenbosch, South Africa
a r t i c l e
i n f o
Article history:
Received 7 July 2011
Received in revised form 14 August 2013
Accepted 22 August 2013
Available online 12 September 2013
JEL classification:
B52
L17
L86
O34
Z13
Z19
a b s t r a c t
We analyze the impact of institutional and cultural factors on the supply side of open
source software (OSS). OSS is a privately provided public good: it is marked by free access
to the software and its source code, and is developed in a public, collaborative manner by
thousands of volunteers as well as profit-seeking firms. Our cross-country study shows that
a culture characterized by interpersonal trust and self-determination/fulfillment values
has a positive impact on OSS activities and the number of developers. The supply side of
OSS also benefits from the enforcement of intellectual property rights. A low degree of
regulation and openness towards scientific progress has a positive impact on the number
of OSS developers, but the latter not on the number of active or core developers.
© 2013 Elsevier B.V. All rights reserved.
Keywords:
Open source software
Institutions
Culture
Social capital
Individualism
Intellectual property rights
1. Introduction
The success of open source software (OSS) has challenged the conventional wisdom on the use of intellectual property
rights (IPRs) and on the private provision of public goods. In the case of OSS, the source code—the human-readable recipe of
a software program—is ‘open’ (disclosed). The OSS licenses grant general access to the software and its source code, as well
as the right to read, modify, improve, redistribute and use it. OSS is developed by a ‘community’ that consists of non-paid
volunteers as well as profit-seeking firms. Nowadays, OSS plays an important role in the ICT sector.1 Thus, OSS is a successful
example of the “private provision of a public good” (Johnson, 2002).
夽 We would like to thank the participants at the Annual Conference of the International Society for New Institutional Economics in Berkeley 2009, the
participants of the Mika Widgrèn Memorial Workshop on Rules, Games and Democracy in Turku 2009, and the participants of the Graduate School ‘The
Economics of Innovative Change’ in Jena for valuable comments and suggestions. In particular we would like to thank Bianka Dettmer, Oliver Kirchkamp,
Florian Noseleit, Maria A. Rossi, Francesco Rullani, Matti Viren, and two anonymous referees. Financial support from the Klaus Tschira Foundation is
gratefully acknowledged.
∗ Corresponding author. Tel.: +49 3641 943 258; fax: +49 3641 943 252.
E-mail addresses: Sebastian.von.Engelhardt@uni-jena.de (S.v. Engelhardt), Andreas.Freytag@uni-jena.de (A. Freytag).
The Apache Webserver software has a remarkable market share, see e.g., the Netcraft Web Server Survey data at www.netcraft.com. Small and mediumsized enterprises and well-known companies like Cisco, IBM, Nokia, Panasonic, Samsung, Sony, and Toshiba use OSS-based business models. Smartphones
running Android and Linux-based ebook readers (e.g., Amazon’s Kindle) are the most prominent examples of OSS-based products.
1
0167-2681/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jebo.2013.08.012
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
91
However, countries differ in terms of their OSS developers per capita as well as in the level of their OSS activity (Engelhardt
et al., 2013; Gonzalez-Barahona et al., 2008), and these differences cannot be satisfactorily explained by GDP or access to
the Internet.
Our study shows that country-specific institutions and culture matters for the willingness to contribute to OSS. Based on
the microeconomics of OSS, we test various hypotheses on how different institutional and cultural factors have an impact
on OSS development. This is a relatively new approach that enables us to deepen the understanding of the phenomenon of
OSS, and helps to clarify some aspects discussed or disputed in the literature. Finally, understanding how country-specific
cultural and institutional factors influence OSS activities is also an important—although so far neglected—aspect for the
discussion about OSS as a concept for developing countries.
We make use of own data on the worldwide allocation of activities of developers registered at SourceForge. We distinguish
between registered users, active users, and core developers and have two measures of activities (postings and software
uploads). We use different strategies to deal with possible country-fixed effects: ordinary least squares (OLS) regressions
with regional dummies, pooled OLS regressions with regional dummies, and first-difference regressions combined with
pooled OLS on the first-difference residuals for the time-invariant variables.
We find that a culture characterized by self-determination/fulfillment favors OSS activities. This is the first empirical result supporting the argument that self-fulfillment motives, which are reported by OSS developers to be important,
are indeed also important for effort, i.e., for OSS activity levels. On the other hand, we cannot support the hypothesis that positive attitudes toward competition foster OSS, which would have been an indicator for the importance of
extrinsic motives for OSS activities. For the number of (core-)developers, however, the importance of extrinsic motives
and incentives that are linked to the existence of an ICT sector find some support, as we see a positive impact of low
economic regulation. Next to self-determination/fulfillment, interpersonal trust is another strong explanatory variable.
Interpersonal trust has a positive impact on the number of OSS (core-)developers as well as on the OSS activity level.
This supports the branch within the OSS governance literature that emphasizes the importance of trust. The notion that
OSS is a kind of anti-IPR (or IPR-less) innovation system is challenged by our findings that the de facto protection (the
enforceability) of IPRs has a positive impact on OSS activities. Finally, openness to novelty—measured as a preference for
new ideas and attitudes towards scientific progress—does not have a significant effect on the number of active or core
developers or OSS contributions. However, scientific progress is significantly correlated with the number of registered
developers.
The remainder of the article is structured as follows. In Section 2, we present the theoretical foundations and derive the
hypotheses for the empirical study. In Section 3 we operationalize the variables, describe the data sources, our sample and
the applied empirical methods. In Section 4, we present and interpret the regression results, before ending with the summary
and outlook in Section 5.
2. Theoretical considerations and hypotheses
In general, cultural and institutional factors shape human interaction and therefore have an impact on the microeconomic level. Hence, in order to derive hypotheses about the influence of institutional and cultural factors on OSS
developers and their activities, we link insights about the microeconomics of OSS with the levels of institutional and cultural
factors.
The different levels that shape economic outcome are, for example, illustrated by Williamson’s, 2000 framework of the
four interrelated levels of social and institutional analysis—see also Fig. 1. Williamson distinguishes between four levels.
The lowest level (Level 4) refers to the focus of neoclassical economics: economic agents maximize their utility, given their
preferences, endowments, payoff-functions, and the rules of the game (Williamson, 2000). Most of the research on OSS
focuses on this level. This comprises contributions that analyze the rationale for firms to develop OSS (e.g., Reisinger et al.,
2013; Llanes and de Elejalde, 2013; Engelhardt, 2010; Henkel, 2006) or the impact of OSS on competition, resource allocation
and welfare (e.g., Bitzer and Schröder, 2007; Economides and Katsamakas, 2006; Mustonen, 2003). Also, research on the
motives of OSS volunteers belongs to Level 4, since it deals with preferences and incentives (e.g., Bitzer et al., 2007; Lakhani
and Wolf, 2005; Ghosh et al., 2002).
On the next level up (Level 3), we find the governance structures within which the Level 4 decisions are made (Williamson,
2000). With respect to OSS, this means that research on the governance structures of OSS projects deals with Level 3 aspects;
including the informal rules (e.g., Langlois and Garzarelli, 2008; Laat, 2007; Markus, 2007; Wendel de Joode et al., 2003),
the way OSS projects use intellectual property law (O’Mahony, 2003), and the role, choice and rationale of the different OSS
licenses (e.g., Sen et al., 2008; Polanski, 2007; Gambardella and Hall, 2006; Lerner and Tirole, 2005).
The next level (Level 2) comprises the “institutional environment” and relates to the definition and enforcement of property rights and contract laws (Williamson, 2000). Clearly, Level 2 directly affects the lower Level 3, since certain governance
structures may not work, may be inefficient or simply not realizable if the higher level institutional environment is not right.
Bad regulations or a lack of enforceability of (intellectual) property rights may make contracts non-enforceable and/or too
costly. With respect to OSS, one relevant aspect of Level 2 is the enforceability of IPRs, since, for example, the OSS licenses
are based on copyright law (Laat, 2005). However, to the best of our knowledge, no study exists that analyzes the impact of
the Level 2 aspects on OSS.
92
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Fig. 1. Williamson’s four interrelated levels of social and institutional analysis.
Source: Williamson (2000), p. 597
The highest level (Level 1, labeled “embeddedness”) is characterized by customs, traditions, norms and religion
(Williamson, 2000); hence, the cultural background of a country. Research on culture and economics has outlined that
culture can affect Level 4 outcomes—i.e., the microeconomic results—via two channels, one direct and one more indirect.
1. Culture influences economic behavior directly, either in the form of social conventions, beliefs (including trust), or individual values and preferences (e.g., Fernández and Fogli, 2009; Guiso et al., 2008; Bisin and Verdier, 2001; Bowles,
1998).
2. Cultural characteristics foster or hinder the implementation and/or functioning of institutions—which in turn, influence
Level 4 (e.g., Tabellini, 2010, 2008; Greif, 1994).
The only study we are aware of that links cultural factors with the geographics of OSS developers is Ramanujam (2007).2
Ramanujam uses data from Ghosh (2006) about the distribution of developers among four world regions and links this with
the four dimensions of culture considered by Hofstede (1991). In his correlation analysis, Ramanujam finds two significant
correlations. The measure ‘Individualism’ (which is related to self-determination and self-fulfillment) is positively correlated
with the number of OSS developers (p-value 0.0683), whereas ‘Power Distance’ (a measure of power inequalities in a society)
is negatively correlated (p-value 0.0086). However, the results should be interpreted with care, as there is no control for
2
His hypothesis is that “cultural differences influence the participation” such that the “[c]ultural differences amongst the programmers from different
regions lead to measurable differences in their participation in the open source movement” (Ramanujam, 2007, p. 16). When interpreting his results, he
gives some plausible explanations for his findings. Nevertheless, the study seems a bit vague in terms of theoretical foundation.
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
93
aspects like number of inhabitants, GDP or Internet access. Furthermore, Ramanujam (2007) distinguishes only four world
regions. Our study runs regressions with data from 61 countries, and analyzes several cultural and institutional factors,
including norms and attitudes.
Lerner and Tirole (2002) formulate probably one of the most famous research questions about OSS: “Why should
thousands of top-notch programmers contribute freely to the provision of a public good?” This refers to the motives of
the OSS volunteers. While Lerner and Tirole (2002) emphasize the role of external motives (peer recognition and signals
for the job market), surveys typically find that intrinsic motives rank higher than extrinsic ones (Lakhani and Wolf, 2005;
Ghosh et al., 2002; Hars and Ou, 2002; Lakhani et al., 2002). However, the surveys reflect what the OSS developers report,
but typically do not link this to the actual effort of the developers or their OSS output. One exception here is Hars and
Ou (2002), who connect the reported motives with individual effort and find that external motives actually have greater
weight—contrary to what the surveys suggest.3 This leads to the question of whether intrinsic or extrinsic motives are more
important. Our study contributes here with respect to internal motives (in this study, self-fulfillment) and external motives.
Hars and Ou (2002) find in particular that self-determination was the strongest intrinsic motive in terms of reporting
(about 80% agreement)—but it had no significant effect on effort. Similarly Hertel et al. (2003) find no effect of “hedonistic
motives” on actual effort. So there seems to be a mismatch between reported motives and their actual relevance, especially
when it comes to ‘self-fulfillment’ motives.4 One possible explanation of this mismatch is that self-fulfillment motives are
important for the decision to participate in OSS (i.e., to become an OSS developer), but not for the actual level of contribution.
Another reason which could explain this mismatch is a possible bias in the surveys. Most of the OSS developers come from
North America and Western Europe (e.g., Gonzalez-Barahona et al., 2008; Lakhani and Wolf, 2005), where an attitude of
self-fulfillment is part of the individualistic culture. So most of the developers surveyed come from cultures where it is
quite common to give self-fulfillment reasons for things one does. The sample bias then translates into a cultural bias,
which can explain both results. Due to the high absolute numbers of respondents from North America and Western Europe,
‘self-fulfillment’ motives rank relatively high but have no effect (too little variance).
Correcting for country size (OSS measures per capita) and distinguishing between membership, roles and contributions,
while varying the cultural background can help to clarify this, since we can make use of the link between culture (Level 1)
and preferences which directly shape Level 4 results. In a culture with a higher emphasis on self-determination and selffulfillment, one would expect that more people engage in individualistic hobbies. And if OSS development is an activity that
fits such preferences, this would also hold for OSS activities. We thus measure the influence of self-determination/fulfillment
within a society on the per-country number of OSS developers and OSS activities. Although Hars and Ou (2002) and
Hertel et al. (2003) find no effects, we do believe that OSS indeed meets values and preferences that are connected to
self-determination and -fulfillment. Therefore we expect the following:
Hypothesis 1. A culture that favors self-determination and -fulfillment has a positive impact on the number of OSS
developers and on the OSS activity level.
Let us now turn to the extrinsic motives, which are—according to Hars and Ou (2002)—more relevant for OSS activities
than what survey reports suggest. One set of the extrinsic motives consists of self-marketing, peer recognition and reputation
within the community (Hars and Ou, 2002; Lakhani et al., 2002). Career aspects define the second set of extrinsic motives,
comprising the improvement of programming skills, i.e., the investment in human capital, and the aim to build up reputation
signals for the job market (Lakhani and Wolf, 2005; Hertel et al., 2003; Ghosh et al., 2002; Hars and Ou, 2002; Lakhani et al.,
2002, in all cases the career motive is stronger than the motives related to peer recognition).
What could explain the mismatch between the reported and the measured relevance of reputation-based incentives?
One possibility is that reputation is just a byproduct: individuals contribute to OSS mainly because of intrinsic motives and
by doing so they get peer recognition and also improve their position in the job market. To gain reputation, however, one
would have to make some kind of significant contribution. This, in turn, implies that those who contribute only a little know
that they do not get the byproduct ‘reputation’, so it is unimportant for them, and they do not mention it. Therefore, if
reputation is more of a byproduct with a certain threshold, then this aspect is irrelevant for those who contribute less, and
therefore Hars and Ou (2002) find a positive correlation between extrinsic motives and effort. Another explanation may
be that—because of the cultural bias discussed above—internal motives (e.g., self-determination) are overestimated in the
survey reports, although the extrinsic motives in the form of peer recognition and career aspects set the incentives, mainly
when it comes to actual effort (and not only the decision to become part of the OSS community). To some degree we can
test for this.
Both sets of extrinsic motives have in common what can be called a “reputation mechanism”. Such a reputation mechanism is an incentive structure based on the merit principle. Furthermore, the relevance of such performance signals indicates
competition, especially when it comes to career concerns. Individuals who believe that competition is good will be in favor
of reputation-based signals. Moreover, individuals who have a positive attitude towards the competition principle will also
3
Hars and Ou (2002) also point out that different types of OSS programmers exist and that, for example, students and hobby programmers are more
internally motivated than professionals.
4
This includes ‘self-determination’, ‘fun’ and the enjoyment of programming work itself or of solving problems, as well as an intellectual challenge—see
Lakhani and Wolf (2005), Hertel et al. (2003), Hars and Ou (2002), and Lakhani et al. (2002).
94
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
be more likely to engage in OSS activities because of “sportive” peer competition as well as for career aspects.5 We can thus
exploit the fact that attitudes towards competition vary over the different countries; hence, cultures.
In a culture with a more positive attitude toward competition, more people will believe that competition is good. An
individual who believes that competition is good is oriented towards external incentives and thus more likely engages in
OSS because of external motives. In addition, positive attitudes towards competition support performance-based reputation.
If more individuals accept the idea of individual performance signals, more peers will be willing to reward good contributions,
and by this support, the functioning of the informal institutions of the OSS reputation mechanisms. This is an example of
how culture in the form of beliefs (Level 1) affects Level 4 (via preference) and the functioning of Level 3 institutions, which
in turn also affects Level 4. Thus, in a country with a more positive attitude towards competition, one is more likely to find
software developers or students who engage in OSS activities because of external motives.
Hypothesis 2. A culture of positive attitudes toward competition has a positive impact on the number of OSS developers
and on the OSS activity level.
Acquiring signals for the job market makes sense only if a relevant job market exists, which implies in the case of OSS the
existence of a hiring ICT sector. Furthermore, OSS enables firms to run OSS business models. Such firms have incentives to
contribute (e.g., Llanes and de Elejalde, 2013; Henkel, 2006). In both cases a growing ICT sector may have a positive impact
on OSS activities. However, world-wide data on each countries’ ICT sectors are scarce. Thus, we cannot directly measure
how the existence of an ICT sector translates into OSS activities via the channels of the external incentives career motives
and business models. However, we can use an indirect measure. Since we control for Internet access, we have a measure
of a country’s ICT level. We now take into account the regulation of economic activities. Strong and distortive regulation
of economic activities in a country has a negative impact on doing business and thus also on the ICT sector. Differences in
countries’ Level 2 aspects—here economic regulation—translate into different degrees of flexibility and performance of the
Level 3 institutions of governance (the costs and benefits of contracts, the setting up of new organizations, etc.), which then
translate into different outcomes on Level 4. This also holds for the ICT sector. We are interested in the question of whether
this effect translates into differences in OSS participation, both on the level of membership and activities. For the incentives
to contribute to OSS, a positive effect would emphasize the importance of career motives and OSS business models.
Hypothesis 3. A high degree of intense economic regulation has a negative impact on the number of OSS developers and
on the OSS activity level.
Without any doubt, extrinsic and intrinsic motives are important; however, motives alone may not tell the full story,
since trust may also play an important role: It is known that social motivations and/or selective incentives (private benefits
of contributing) can transform a prisoner’s dilemma into a coordination game (Sen, 1974; Ullmann-Margalit, 1977; Rabin,
1993; Fehr and Schmidt, 1999; Bicchieri, 2006). So given strong enough extrinsic and intrinsic motivations, OSS production
will be a coordination game, where multiple equilibria exist (Hippel and Krogh, 2003; Osterloh and Rota, 2004). Which
equilibrium actually establishes depends on the players’ expectations about the contributions of the other players; in other
words, the contribution decision will depend on trust (Osterloh and Rota, 2004). And indeed, the literature on public good
problems indicates that interpersonal trust has a positive impact on cooperation and reciprocal behavior (Yamagishi et al.,
2005; de Cremer, 1999; Ostrom, 1998; Yamagishi and Yamagishi, 1994). This leads to the trust vs monitoring and control
question (Markus, 2007) in the research on OSS governance structures, where some researchers see trust to be essential for
OSS development, while others oppose this view. So the current literature gives different, and even contradicting answers.
Sharma et al. (2002) see a high level of mutual trust being necessary and a key element, distinguishing OSS communities
(based on trust) from traditional organizations (not based on trust). Or as Bergquist and Ljungberg (2001) put it: “Virtual
collaboration [here: OSS projects] puts high demands on people having trust in one another.” Laat (2010) argues that the
hacker culture (which existed before) was important in developing a climate of trust in OSS. Here, general trust (which
existed ex ante) was important to establish the more specific, i.e., personal trust, among the members of the diverse OSS
projects.
Further research emphasizes the complementary relation between trust and control and analyzes how informal and
formal rules (including licenses) can improve and sustain trust (Osterloh and Rota, 2004; De Noni et al., 2013). Since actual
behavior in OSS projects can easily be monitored, developers can build credibility and reputation based on their track records,
which then also determine whether developers are assigned to more important tasks and have certain decision rights, e.g., by
becoming core developers (Sinha et al., 2011; O’Mahony and Ferraro, 2007, 2012). Clearly, the better the control mechanisms,
like monitoring, reputation and track record-based ‘careers’ work, the less the OSS governance would have to rely on trust.
Gallivan (2001) claims that trust is not a necessary condition for OSS projects, since its control mechanisms ensure effective
performance. Such a view is supported by Hertel et al.’s, 2003 empirical work, finding that trust plays only a minor role in
OSS activities. However, this is contrasted by Stewart and Gosain (2006), who find a positive impact of trust on OSS team size
and team effort. Moreover, Hertel et al.’s, 2003 results may be driven by sample bias (Hertel et al., 2003, p. 1175), since the
5
Notice that we state a logically sufficient but not necessary condition here. Not all who contribute because of external motives will prefer competition.
But those who think that competition is good will respond to external motives.
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
95
authors analyze participants’ engagement in one OSS project only (the Linux kernel, the core of the Linux operating system)
and membership in this development group is determined by a good track record and reputation.
So the literature expresses two different views, and it is remarkable that the only two studies that link trust to effort
have contradicting findings. We can contribute to the research on the role of trust in the governance of OSS by analyzing the
impact of general trust on OSS activities. Social capital (Putnam, 1995, 1993) in terms of interpersonal trust6 varies across
different countries. We can exploit this fact. The different cultures of the countries (Level 1) translate into different beliefs (in
this study, trust), which in turn directly affect the outcome of a given microeconomic situation. This enables us to measure
the importance of general interpersonal trust for OSS. If trust is essential, then having more general trust makes it easier to
establish the (more specific) trust within the various OSS projects as well as to sustain and enhance it with complementary
governance rules. In other words, if trust is essential, we would expect a positive impact of general trust on the number of
OSS developers and OSS activities; otherwise, if trust is not important, we would expect to see no impact of general trust.
For theoretical reasons and the quite possible sample bias problem of the Hertel et al. (2003) study, we assume that trust
plays a role.
Hypothesis 4.
Interpersonal trust has a positive impact on the number of OSS developers and on the OSS activity level.
An important formal institution of the OSS governance mechanisms that can also support and stabilize trust is the
definition of property rights, in our case OSS licenses (De Noni et al., 2013). Licenses like the GPL7 safeguard the commonly
developed OSS against the risk of opportunistic exploitation (i.e., turning OSS code into proprietary software) by applying a
“copyleft” clause, which ensures that the open source code stays open source.8 Thanks to its legal structure, the GPL aims to
ensure that egoistic motives do not crowd out altruistic motives (Franck and Jungwirth, 2003). OSS licenses can fulfill this
task only if they are legally binding. They are, like software licenses in general, based on copyright law (Laat, 2005) which
grants copyright to the creator of the code. Based on these enforceable IPRs, the code owner can define the inclusive OSS
license to open up the code and secure its openness with the “copyleft” clause. There exist a variety of different OSS license
types (see e.g., Lerner and Tirole, 2005), differing in how they restrict the usage of the code. In particular, firms ‘owning’ OSS
projects make use of sophisticated licenses and dual-license strategies (e.g., Välimäki, 2003), as it is crucial for them to define
exactly what is exclusively owned and what not. Obviously, such legal arrangements are only possible and effective if IPRs are
respected and such licenses can be enforced.9 Furthermore, OSS governance structures also indirectly rely on trademarks.
The core developers of an OSS project control the project by using their exclusive rights to decide whether to accept or
reject contributions. These passive control rights are enforced through legal control of the project’s trademarked name (and
by controlling access to the database in which the software is stored) (Engelhardt, 2008; Wendel de Joode et al., 2003;
McGowan, 2001). Trademark protection prevents the cloning of projects and supports the signaling function of the project’s
name. Here, protection of IPRs indirectly supports the OSS governance structures and the informal institution ‘reputation’.
Thus, both the non-commercial part of the OSS community as well as the firms involved benefit in practice from the prospect
of defining and enforcing IPRs. O’Mahony (2003) shows in detail how OSS projects use intellectual property law to protect
their work.
Therefore, one could conclude that OSS depends on the enforceability of license terms, and hence, on the enforceability of
IPRs. However, some quarters within the OSS movement argue against intellectual property: the Free Software Foundation
opposes the use of the term “intellectual property”, and Richard Stallman refuses the idea of intellectual property.10 Beyond
the OSS community, OSS is sometimes confused with ‘software without IPRs’, and OSS and intellectual property are perceived
as being opposites (Lindberg, 2008; Fitzgerald, 2006).11 Furthermore, the case of OSS is sometimes used to argue against IPR
protection, for example, by NGOs (Lewis, 2008).
Our cross-country study can help to clarify this point. Countries differ with respect to the de facto enforcement of IPRs,
which belongs to the institutional environment (Level 2). This has a direct impact on the lower level “governance”—in our
context, the governance of OSS in term of the enforcement of the formal institution ‘OSS license’—which in turn shapes the
outcome of the Level 4 game—in our case, the OSS activities. If, as we assume, OSS licenses (and thus OSS governance) benefit
from the enforceability (the security) of IPRs, a better (weaker) enforceability of IPRs should translate into more (less) OSS
activities.
6
Social capital “refers to features of social organization such as networks, norms, and social trust that facilitate coordination and cooperation for mutual
benefit” (Putnam, 1995, p. 67). Some scholars refer to the number of ties only, others stress the features, strength and quality of such ties, which includes
trust. We focus on trust in this study.
7
GPL stands for the GNU General Public License.
8
A copyleft clause states that any further developed software as well any derived work must be licensed as a whole under the same OSS license as the
original code.
9
See e.g., Kumar (2006) on the GPL; for current examples of “the GPL in court” visit the webpage http://gpl-violations.org.
10
This view is opposed by, e.g., Eric S. Raymond, co-founder of the Open Source Initiative. Raymond supports the idea of property right claims, and hence,
also of IPRs, but argues that proprietary software is simply an inefficient way of developing software (see Weber, 2004a). Others, like Greg Perkins, also
point out that “Open Source depends on the idea of the individual human right to private property” (Perkins, 1998).
11
See, for example, the notion of OSS as anti-IPR in Paun et al. (2012), Kimppa (2008), and Dafermos and Sderberg (2009), who state that OSS shows that
software can be produced “without intellectual property relations”.
96
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Hypothesis 5. The protection and enforceability of IPRs has a positive impact on the number of OSS developers and on the
OSS activity level.
Our final argument—the influence of openness to novelty—is again an example of how culture (Level 1) affects preferences
and values, which in turn, shape Level 4 results. We argue that in societies that are more open to novelty, a higher share
of people embrace new ideas and scientific progress, which in turn shapes the likelihood of inhabitants choosing certain
activities, here: OSS development.
A preference for new ideas is a good precondition for the adoption of the OSS model and for active participation, for
two reasons. First, the process of (open source) software development is a search for new solutions, i.e., an innovative
process as such. And OSS development is a new way of developing novelty; OSS governance, a new way of “coordinating
innovation”, (Kugler, 2005) based on a new IPR paradigm (Maurer and Scotchmer, 2006), sometimes even perceived as
an “intellectual property revolution” (Pisano, 2006). OSS creates innovation and at the same time is an innovation at the
level of how to organize software development. Second, some authors state that OSS leads to more innovation than the
“traditional” proprietary method (Piva et al., 2012; Rossi-Lamastra, 2009; Ghosh, 2006), which could be because OSS offers
better conditions for innovation processes (see e.g., Rossi-Lamastra, 2009) but also because OSS attracts more innovative
actors, i.e., actors that are more open to novelty, or both.
Neither the literature on the innovativeness of OSS nor that on the adoption of OSS in developing countries (e.g., Yildirim
and Ansal, 2011; Câmara and Fonseca, 2007; May, 2006) has so far taken into account cultural aspects like attitude towards
novelty. If openness to novelty is indeed a determinant—or even a precondition—for participation in OSS, then this would
be a new argument for the view that OSS spurs innovation and an important information for the discussion on the adoption
of OSS strategies for developing countries. The same holds for the second aspect of openness towards novelty: the positive
attitude towards scientific progress.
Beyond the fact that a positive attitude towards scientific progress also measures openness towards novelty, openness
towards scientific progress is a good precondition for the adoption of the OSS model and for active participation, for the
following two reasons. First, OSS itself has a strong technical aspect. It is a novelty from ‘cyberspace’, connected to programming, computers and the Internet, and therefore, to the technical aspect of scientific progress. Second, historically, the
OSS movement stems from software science and is therefore rooted in scientific culture. The best known OSS license, the
GPL, was developed by Richard Stallman, who worked in the MIT Artificial Intelligence Lab. As for its culture of discussing
problems in online forums, its openness and reputation mechanism, similarities to open science can be observed (Dalle and
David, 2005; Lerner and Tirole, 2002; Giuri et al., 2002). We thus argue that a pro-science culture—that translates into proscience attitudes, i.e., preferences and values—may support the adoption of the cyberspace-located OSS model, including its
science-like formal and informal institutions. Since a positive attitude towards scientific progress is a sufficient (albeit not
necessary) condition for pro-science preferences, we expect a positive impact on OSS.
So we expect that it is more likely to find (active) OSS developers in societies with a culture more open to novelty.
Hypothesis 6.
level.
A preference for new ideas has a positive impact on the number of OSS developers and on the OSS activity
Hypothesis 7. A positive attitude towards scientific progress has a positive impact on the number of OSS developers and
on the OSS activity level.
In addition, some control variables are necessary. The data on the geographic origin of OSS developers show that most
OSS contributions come from developed countries. Therefore, we control for GDP per capita. Furthermore, we control for
education, because studies like Ghosh et al. (2002) indicate that OSS developers are well-educated software engineers or
ICT students. Finally, without Internet access there is no access to the online community of OSS developers.
3. Data and method
3.1. Data on OSS developers registered at SourceForge
SourceForge is the largest site hosting OSS projects12 and an often-used source for research on OSS.13 We make use of data
offered by the SourceForge Research Data Archive (SRDA).14 This archive comprises of monthly dumps which contain some
of the information stored at SourceForge. We use SRDA to extract information on the developers’ countries, their activities
and roles.
Here, we give a brief description of how we derived information on the geographical origin of the developers; for more
details, please refer to Engelhardt et al. (2013). In registering with SourceForge, users have to provide a valid email address.
12
SourceForge is an Internet platform, i.e., a virtual center, where the developers of OSS projects can discuss, coordinate their tasks, upload code, etc.
While access to the developer areas requires registration, the software can be downloaded by anybody.
13
See e.g., Giuri et al. (2010), David and Rullani (2008), Gonzalez-Barahona et al. (2008), Fershtman and Gandal (2008), Lerner et al. (2006), and Lerner
and Tirole (2005).
14
SRDA is offered by the University of Notre Dame under a special agreement for scientific research.
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
97
Where the email address ends with a country-coded top-level domain (TLD), one can identify the user’s country, as in the
case of.uk. In the case of a generic TLD like.com or.org, we use the second-level domain (SLD) to locate the server of the email
service provider. Given that the developer has indicated a well-defined and unique time zone, this can be used to assign
a country to a user (e.g., “Europe/Berlin” implies Germany). And finally, some of the monthly SRDA dumps contain the IP
addresses of some developers. With an IP address it is possible to locate the developer (the respective server of the Internet
service provider used) via GeoIP. Since IP, time zone and country-coded TLD are more reliable than the SLD-based approach
(which is also reflected by the matching rates we get from cross-checking the methods—see Engelhardt et al., 2013), the data
set of our study was derived by the following method. Developers’ geographic locations were primarily identified by the IP
addresses assigned to them. Developers whose IP addresses were not stored in the data were located by the country-coded
TLDs of their email. In the case of emails with generic TLDs, we used time zone information if applicable; otherwise the SLD
approach. The SRDA monthly coverage starts in 11/2004, and the latest data available for our purpose are for the year 2006.
OSS developers are heterogeneous, both with respect to their roles (core developers vs bug fixers, etc.) and with respect
to their activities, such as, code contributions (see e.g., David and Rullani, 2008). We are able to assign to each user the
number of posted messages and uploaded software files. This enables us to distinguish active developers (those who posted
messages or uploaded files in the respective time frame) from non-active ones. Furthermore, we are able to identify whether a
developer was an officially assigned developer or an administrator of a project. We will refer to the group of officially assigned
developers and administrators as the ‘core developers’.
Once registered, a developer’s account stays in the SourceForge database, and since the data provided by SRDA contain
UNIX timestamps, we can identify the date of registration. So we are able to locate 94% of all developers who had registered
up to the end of 2006, and for all of these approximately 1.4 million developers, we know when they registered. This enables
us, for example, to compare the per-country number of registered users for 2004 with those for 2006.
The time period for which we have information on the activities and roles of the developers is more limited. Activities
and roles do change and, as mentioned before, SRDA monthly coverage starts not earlier than in 11/2004. The monthly
dumps of the SRDA do however contain data from previous months (for example, the 11/2004 dump also contains data with
timestamps of 10/2004, 9/2004, and so on) but these old entries are incomplete. Comparing ‘old’ data of the monthly dumps
of 12/2006 and earlier with the actual ‘old’ dumps revealed that, on average, there is a loss of 5.6% of data of minus one
month (i.e., the May 2005 data, for example, are expected to contain 94.4% of the April 2005 data). Therefore, we decided
that we could use the information about activities and roles stored in the 11/2004 dumps for 10/2004, but not for earlier
months.
Weighting all of this information by the number of inhabitants enables us to calculate country-specific information.
Comparing our data with Gonzalez-Barahona et al. (2008) we find similar patterns of geographical distribution (see also the
world maps of OSS developers and activity levels in Engelhardt et al., 2013). As we have information on the level of OSS activity
(measured in files uploaded or messages posted per population) and the number of OSS developers per population (where
we distinguish between core developers, active developers and registered developers) our data offer more information on
global OSS activities than any other non-survey data we are aware of.
3.2. Data on GDP, education and ICT
We take into account the GDP per 100,000 inhabitants for 2003 and 2005, in purchasing power parity (World Bank, 2007).
To weight our country-specific OSS measures by the number of inhabitants, we use population data offered by World Bank
(2007). As a measure for education, we use the UNDP Education Index 2003 and 2005—for robustness checks we also use
the gross enrollment ratio for tertiary schools and the combined gross enrollment ratio for primary, secondary and tertiary
schools (UNDP, 2007, 2005). Worldwide data on, e.g., the number of software developers, the size of the software sector, or
other differentiated data on the ICT sector are poor. The best data available refer to Internet access.15 We use the number
of Internet users per inhabitants (‘Internet users’) of 2003 and 2005 as a proxy here. The data for this come from ITU (2005,
2003).
3.3. Data on regulation and IPR protection
In order to evaluate the degree of business regulation, we use the sub-index “5 Regulation of Business” of the Economic
Freedom of the World Index (Gwartney et al., 2008). We use the data for 2003 and 2005, and denote this variable by
‘regulation’.16
With respect to IPR, we use one of the sub-indices of Gwartney et al. (2006): the sub-index of the protection of IPRs
(“2C Protection of intellectual property”), which is based on data from the Executive Opinion Survey of the World Economic
15
At least for some countries, data on the share of employees working in the ICT sector are available. But with the share of ICT employees and Internet
users, we would run into problems of multicollinearity since each Internet access must have been installed by someone working in the ICT sector. Therefore,
we leave ICT employees out and use only Internet use. We also do not to use real prices of ICT, because of a lack of data.
16
The original measure indexes deregulation. To avoid confusion, we use an inverse version, so that highly regulated countries have high scores.
98
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Forum.17 Since the latest data on IPR are for the year 2004, we use the index values of 2002 and 2004. We denote this index
by ‘IPR protection’.
3.4. Data on cultural factors
Culture changes very slowly (Triandis and Gelfand, 2012; Hofstede, 2001; Williamson, 2000); therefore, for the time
frame of our analysis, cultural factors are time-invariant. Our main source of our cultural factors is the World Values Survey
(WVS). The WVS offers a wide range of country-specific cultural data and is often used in cross-cultural research; examples
include international studies and global affairs (Abdollahian et al., 2012; Inglehart, 2000), cross-cultural psychology (Welzel,
2010; Inglehart et al., 2008; Diener et al., 2000; Inglehart and Klingemann, 2003), political science (Inglehart and Welzel,
2010; Inglehart, 2008; Welzel et al., 2005), sociology (Breznau et al., 2011; Sandholtz and Taagepera, 2005; Nevitte and Kanji,
2002; Norris and Inglehart, 2002; Inglehart and Baker, 2000; Nicolás, 1996), and economics (Alesina et al., 2013; Marini,
2013; Johnson and Mislin, 2012; Dearmon and Grier, 2009; Guiso et al., 2008; Fisman and Khanna, 1999).
We use the latest country data available for each. The WVS is carried out in several waves. Although most countries are
covered in all waves, some are not. For each variable and country, we select the survey data for the most recent year using
WVS (2010) and EWVS (2006). The majority of country data come from surveys undertaken between 2000 and 2006; only
a few date back to earlier years.18
3.4.1. Degree of self-determination/fulfillment
We make use of several variables available that measure the dimension of culture which are linked to self-determination
and self-fulfillment.
First, we use a measure of ‘individualism’ in the tradition of Hofstede (1991), a concept that focuses on rights above
duties, concern for oneself (and one’s close family) and self-fulfillment. Hofstede developed his original individualism index
for 50 countries based on a worldwide survey of IBM employees that was carried out during 1978-83. The questions the
individualism index was built upon covered whether the job leaves sufficient time for personal and family life, considerable
freedom to adopt own approaches, includes challenging work, offers opportunities to improve and learn new skills, etc.
Hofstede, 1991, p. 49ff). Based on these categories, high scores for individualism indicate the prevalence of individual interest
in a society, implying that people would like to—and can—“do what they want to do”, i.e., individualism emphasizes personal
autonomy and self-fulfillment. We use an updated and further-developed version of Hofstede’s measure, namely, a merging
of ratings provided by Triandis’ and Hofstede’s scores, as used by, for example, Suh et al. (1998), Diener et al. (2000) and
Oishi (2000).
We also take into account a measure of the ‘importance of leisure time’ from the WVS, since this captures one of the
original items that Hofstede’ s individualism measure was built upon (Hofstede, 1991). The WVS question asks how important
leisure time is in one’s life. Possible answers are given in a four-point scale ranging from “Very important” to “Not at all
important”.
Next, we also consider individual responsibility, because individualistic cultures emphasize that it is the individual who
is responsible—not the collective (Triandis, 1995). Indeed, self-responsibility belongs to individualistic values (Shulruf et al.,
2007; Realo et al., 2002) and individualism is strongly linked to personal responsibility (Kemmelmeier et al., 2006; Basabe
and Ros, 2005; Allik and Realo, 2004; Furrer et al., 2000). We use two measures here. The first is the importance of selfresponsibility (the question asks one to place oneself on a range of 1–10 in terms of expressing one’s own opinion, with
1 = People should take more responsibility to provide for themselves vs 10 = The government should take more responsibility
to ensure that everyone is provided for). Second, the WVS delivers the percentage of all respondents for a country who
mentioned that a “feeling of responsibility” is an important quality children should learn at home. (They are given a list of
qualities that children could be encouraged to learn at home. They are to choose up to five they considered to be especially
important.)
Another important aspect in this context is Inglehart’s concept of ‘post-materialism’ which refers—because of the way it
is defined—to how much people emphasize self-expression, self-determination, freedom of speech, quality of life, etc., over
more materialistic aspects like physical sustenance and safety or the maintenance of order in a society (Inglehart, 1990,
1981, 1977). A more post-material culture—in the sense of Inglehart’s definition—is more in favor of self-determination
and self-fulfillment. Moreover, post-materialism is strongly correlated with individualism (Brons, 2006; Basabe and Ros,
2005), which may not be surprising since post-materialism is one important aspect of the higher-level cultural dimension
‘self-expression’ (Inglehart, 2006; Inglehart and Oyserman, 2004), which in turn, taps the same dimension of culture that
cross-cultural psychologists refer to as individualism, measured by Hofstede and Triandis (Inglehart, 2006; Inglehart and
Oyserman, 2004).
17
In the survey the question is: IPR protection “in your country, is 1 = weak and not enforced, up to 7 = strong and enforced”. However, the data published
by Gwartney et al. (2006) are recoded to a wider scale.
18
For example, 65.57% of our data on interpersonal trust are from 2000 to 2006 while 90.16% are from 1999 to 2006; 63.33% of the data on the attitude
towards scientific progress are from 2000 to 2006 and 93.44% are from 1996 to 2006. The oldest data we use are those on the preferences for new ideas:
70.49% of these data are from 1996 to 2006, hence 29.51% are from 1990 to 1995. In all cases the oldest entries are from 1990.
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
99
We want to have one single measure for the degree to which the culture of a society favors self-determination and selfexpression. We thus apply principal component analysis (PCA) to our various measures. We construct a principal component
labeled ‘self-det/fulfill’, consisting of the measures described above. The Kaiser–Meyer–Olkin measure of sampling adequacy
of ‘self-det/fulfill’ is 0.77. Note that our measure of ‘individualism’ was limited in the number of countries available (data
were available for only 56 countries), while the other measures were available for 89 and 88 countries respectively. In order
to avoid an overly limited dataset, we choose the option of the PCA to construct the principal component with missing data
replaced by the mean. As a robustness check, we also run regressions with disaggregated data—we will return to this when
we discuss our results.
3.4.2. Attitudes toward competition and the merit principle
For the degree of positive attitudes toward competition, we use the corresponding question in the WVS. The survey
question asks one to place oneself according to the range of opinion specified by “Competition is good. It stimulates people
to work hard and develop new ideas” vs “Competition is harmful. It brings out the worst in people”.19
3.4.3. Interpersonal trust
Interpersonal trust in the WVS survey is covered by the question “Generally speaking, would you say that most people
can be trusted or that you need to be very careful in dealing with people?” The potential answers are “Most people can be
trusted” vs “Can’t be too careful” and “Dont know”. We use the percentage of respondents answering “Most people can be
trusted” and label this variable ‘trust’.
3.4.4. Attitudes towards novelty (new ideas and scientific progress)
Data on attitudes towards novelty, namely, on a preference for new ideas and scientific progress also come from the WVS.
Our first measure is ‘prefer new ideas’. The WVS asks one to rate oneself on a scale specified by 1 = “Ideas that stood test of
time are generally best”, up to 10 = “New ideas are generally better than old ones”. Our second measure is ‘science adv help’,
which represents the percentage of respondents saying that scientific advances will help mankind (The question is “In the
long run, do you think the scientific advances we are making will help or harm mankind?” Possible answers are: 1 Will help,
2 Will harm, 3 Some of each).
3.5. Empirical method and sample
There may be some unobservable country-specific characteristics or regional effects. We use the following strategies to
deal with this problem. First, we include regional dummies. Second, in addition to OLS, we also run pooled OLS (POLS) and
first-difference (FD) regressions. However, our cultural variables are time-invariant (at least given our time horizon) and are
thus dropped from the FD regressions. To deal with this problem, we apply the same strategy as Boyce and Wood (2011):
we first apply an FD estimation and then run POLS on the FD residuals for the time-invariant cultural variables.
For the POLS and FD estimations, we need two time periods. Our data on OSS activities and roles range from 10/2004 to
12/2006. We use two six-month periods: the first period (t = 1) with OSS data from 10/2004 to 3/2005 and the second period
(t = 2) with OSS data from 07/2006 to 12/2006.20
We run regressions of our different OSS measures on seven explanatory variables representing our hypotheses plus three
control variables. As the data for the cultural and institutional variables and our control variables are not available for all
the countries, we finally obtain a sample with 61 countries of which four countries have missing data on Internet access,
IPR protection and business regulation for t = 1 (Table A.1 in Appendix A). Table 1 provides the descriptive statistics for our
two-period dataset and for correlations among the regressors, see Table 2. Table A.2 in Appendix A lists the definitions and
data sources of the regressors.
Our different OSS measures are positively skewed, so we use log-linear models21 —with log(1+[OSS measure]). Thus,
for FD the regressors explain the percentage change in the respective OSS measure. For each of our different logged OSS
measures, we run and compare (a) a simple OLS regression using the latest period only, (b) a POLS regression and (c) a
combination of FD regression followed by a POLS regression on the FD residuals for the time-invariant variables.
4. Empirical results
In general, the results appear quite robust. In this section we present and describe the most representative models
(Tables 3 and 4). In addition, Table 5 provides an overview of the results of further variations. We refer to this further on
where we discuss the results. Table 3 presents the regression results for the logged OSS activities, measured as the number
of uploaded files and posts, each per 10,000 inhabitants. In a second step, we run regressions of the logged number of OSS
19
Scale with 1 = Competition is good to 10 = Competition is harmful—we recoded this such that high values indicate a positive attitudes toward competition.
Since the year 2004 is not fully covered by our OSS data we cannot split the data into three periods. To have more within variance we therefore pick
two periods with a gap.
21
Only GDP is logged as well.
20
100
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Table 1
Descriptive statistics.
Variable
Mean
Time-variant variables
log(GDP)
overall
between
within
education
overall
between
within
Internet users
overall
between
within
regulation
overall
between
within
IPR protection
overall
between
within
Time-invariant variables
self-det/fulfill
competition
trust
prefer new ideas
science adv help
Std. Dev.
Min
Max
Observations
2.492365
0.9716144
0.9736083
0.0483449
−0.1553675
−0.1242452
2.376581
3.863385
3.837542
2.60815
N = 118
n = 61
T-bar = 1.93443
0.8958475
0.1193668
0.1179348
0.0078174
0.44
0.455
0.8658475
0.99
0.99
0.9258475
N = 118
n = 61
T-bar = 1.93443
0.3307701
0.2602868
0.2585931
0.0415691
0.0016388
0.0020276
0.235684
0.8700264
0.8510047
0.4258563
N = 118
n = 61
T-bar = 1.93443
4.296782
1.292188
1.186965
0.5235978
0.985198
1.681885
3.249806
7.433333
7.159005
5.343758
N = 118
n = 61
T-bar = 1.93443
5.036251
2.166877
2.183521
0.2631744
1.810036
1.810036
4.24571
8.970276
8.818471
5.826791
N = 118
n = 61
T-bar = 1.93443
0
3.76319
0.2745596
5.319818
0.4176392
1.687137
0.5988677
0.1540374
0.8356821
0.1797441
−4.895826
2.491
0.0477969
3.630121
0.1306413
3.200135
5.027054
0.7416503
8.672649
0.8174847
61
61
61
61
61
Note: For the time-variant variables we report panel descriptives while for the time-invariant we report t = 2 where all 61 countries are covered.
developers, where we distinguish among the number of core developers, the number of active developers and the number of
registered developers (Table 4). As with the OSS activity measures, the OSS developer measures are all per 10,000 inhabitants.
Both tables are structured in the same way. After the control variables, the influence of the explanatory variables presenting
our hypotheses is shown (we put time-variant variables first). For each OSS measure, we present the results of OLS, POLS,
and FD followed by the POLS regressions on the residuals of the FD estimates.
In all cases we run the regressions with (cluster-)robust standard errors, i.e., heteroskedastic- and autocorrelationconsistent estimates. Furthermore, we check for potential problems with multicollinearity, by looking at the pairwise
Table 2
Correlation table.
log(GDP) education Internet
users
log(GDP)
education
Internet users
regulation
IPR protection
self-det/fulfill
competition
trust
prefer new ideas
science adv help
regional dummies
Africa
America, North
America, South
Asia
Europe, Central
and East
Europe, Western
Middle East
Scandinavia
regulation IPR
protection
self-det/
fulfill
competition trust
prefer new
ideas
science adv
help
1
0.815
0.847
1
0.633
−0.646
0.801
0.733
0.330
0.498
−0.215
−0.365
−0.466
0.551
0.545
0.365
0.325
−0.347
−0.243
−0.784
0.853
0.691
0.136
0.694
−0.167
−0.235
1
−0.763
−0.595
−0.0590
−0.546
0.125
0.201
1
0.748
0.225
0.629
−0.147
−0.277
1
0.207
0.484
0.123
−0.359
1
−0.0261
−0.142
−0.265
1
−0.142
−0.0893
−0.397
0.180
−0.230
−0.336
−0.0419
−0.352
0.0762
−0.146
−0.359
0.197
−0.264
0.145
−0.350
−0.144
−0.132
0.114
−0.0890
0.210
0.144
0.195
−0.105
0.193
−0.301
−0.145
−0.337
−0.201
0.209
0.0411
−0.413
−0.371
−0.272
−0.105
0.194
−0.138
−0.0920
−0.157
0.0510
0.0757
0.0826
−0.333
0.419
0.129
−0.0571
−0.238
−0.382
0.256
0.0219
−0.213
0.146
0.0275
0.470
−0.0169
0.327
0.274
−0.0893
0.233
0.320
−0.0972
0.547
−0.225
0.0153
−0.483
0.486
−0.110
0.407
0.451
−0.0277
0.436
0.309
0.0940
−0.124
0.0844 −0.00163
−0.193
−0.00539
0.678
0.00834
−0.200
0.277
−0.0103
1
Africa is the residual dummy skipped in the regressions.
1
0.0999
1
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
101
Table 3
OSS activities.
Files (per 10,000 inhabitants)
OLS
−0.264
(−0.87)
−0.246
education
(−0.15)
0.740
Internet users
(0.98)
regulation
−0.498
(−0.37)
1.761
IPR protection
(1.53)
2.790***
self-det/fulfill
(2.99)
−1.668
competition
(−0.95)
trust
2.105**
(2.13)
prefer new ideas
−0.658
(−0.60)
0.280
science adv help
(0.49)
(regional dummies skipped)
Constant
0.0658
(0.40)
log(GDP)
Observations
Adjusted R2
61
0.742
Posts (per 10,000 inhabitants)
POLS
FD
−0.324
(−1.46)
−0.392
(−0.32)
0.967
(1.32)
−0.0536
(−0.05)
2.820***
(2.76)
2.404***
(2.76)
−1.333
(−0.92)
1.390*
(1.82)
−0.399
(−0.42)
0.181
(0.41)
−0.930
(−1.03)
1.643
(0.63)
0.582
(0.40)
−0.386
(−0.45)
1.847
(1.22)
POLS (resid)
OLS
POLS
FD
−0.962
(−1.11)
2.683
(0.53)
2.941
(1.07)
−1.965
(−0.73)
9.094***
(4.16)
8.482**
(2.24)
−1.952
(−0.39)
12.13***
(4.17)
3.548
(1.43)
−0.570
(−0.39)
3.772
(1.19)
19.16*
(1.77)
−5.150
(−1.31)
3.541
(0.99)
0.682
(0.17)
3.962***
(3.29)
−0.452
(−0.29)
2.549***
(2.88)
−0.718
(−0.64)
−0.347
(−0.59)
−0.440
(−0.38)
3.489
(0.54)
1.672
(0.50)
−6.880*
(−1.68)
5.885*
(1.69)
7.539*
(1.78)
−0.593
(−0.10)
12.31***
(3.86)
4.079
(1.27)
0.0900
(0.05)
0.00688
(0.06)
118
0.773
POLS (resid)
0.0884
(0.24)
−0.0945
(−0.96)
−0.404
(−0.63)
−0.534
(−1.27)
−2.165
(−1.50)
−0.0275
(−0.04)
118
0.003
(within)
118
0.796
61
0.804
118
0.845
118
0.076
(within)
118
0.398
1.999
(0.23)
−13.04
(−1.56)
12.57**
(2.56)
11.74
(1.34)
−0.993
(−0.29)
t statistics in parentheses.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
correlations and also by checking the Variance Inflation Factors (VIFs). Table A.3 in Appendix A reports the VIFs. ‘IPR
protection’ is the only explanatory variable22 that has VIFs above 10—albeit for OLS only. We will come back to this.
Comparing the results over all tables shows that OSS is mainly influenced by the degree of self-determination and
self-fulfillment, interpersonal trust, the protection and enforcement of IPRs, and low economic regulation. The variables
‘self-det/fulfill’ and ‘trust’ are significant for all OSS measures, except the number of registered users. ‘IPR protection’ is
significant in all POLS and all but one OLS regressions, while ‘regulation’ is significant in all POLS and FD regressions of all
three groups of OSS developers.23 A positive attitude towards competition has no significant impact. While a preference for
new ideas it not significant at all, a positive attitude towards scientific progress is significant in regressions of the number
of registered users.
So the significant positive correlation of OSS with both ‘self-det/fulfill’ and ‘trust’ is a robust result (see also the variation
Table 5). Therefore, we neither reject Hypothesis 1 nor Hypothesis 4. It has been argued that self-determination/fulfillment
values–in particular if expressed by individualism—is at odds with high social capital, here: interpersonal trust. However,
there are good theoretical reasons to rather assume a positive relationship.24 Indeed empirical findings support such a view.
More individualism goes hand in hand with more tolerance and trust (Hofstede, 2001); several studies show that there is a
positive relationship between individualism and trust (Li and Fung, 2013; Berigan and Irwin, 2011; Realo et al., 2008; Allik
and Realo, 2004) or pro-social behavior towards strangers, such as, charitable giving and volunteerism (Kemmelmeier et al.,
2006). Moreover, post-communist societies have less interpersonal trust (Bjørnskov, 2007). We also run regressions on the
limited sample of countries were the Hofstede and Triandis individualism measure was available with (a) this individualism
22
The control variable log(GDP) has VIFs above 10 because of the regional dummies (without the dummies the VIFs are below 10).
It is also significant in the OLS for posted messages.
The theoretical argument is that spontaneous and voluntarily established links with each other as well as voluntary cooperation—also with
strangers—favor generalized trust. This line of thought dates back to, e.g., Durkheim (1933), who argues that the division of labor unites peoples since
it increases the need to coordinate. Collectivist countries have more hierarchical and in-group vs out-group-related social relationships, which imply
more social control and sanctions and preferences for members of the (relatively small) in-group. This undermines general trust (Yamagishi et al., 1998;
Yamagishi and Yamagishi, 1994). For example, cheating may be acceptable if (but only if) the target is not a member of the respective in-group (Triandis
and Gelfand, 2012).
23
24
102
Table 4
OSS developers.
OLS
−0.649
(−0.59)
0.676
education
(0.11)
5.388*
Internet users
(1.73)
regulation
−1.850
(−0.48)
7.142**
IPR protection
(2.05)
8.438**
self-det/fulfill
(2.34)
−0.801
competition
(−0.14)
7.524***
trust
(2.88)
0.949
prefer new ideas
(0.31)
science adv help
1.471
(0.86)
(regional dummies skipped)
−0.303
Constant
(−0.53)
log(GDP)
Observations
Adjusted R2
61
0.852
t statistics in parentheses.
*
p < 0.10.
**
p < 0.05.
***
p < 0.01.
POLS
FD
−0.700
(−0.86)
−1.095
(−0.26)
6.314**
(2.56)
−5.968**
(−2.46)
4.951**
(2.47)
7.519**
(2.46)
0.244
(0.05)
6.888***
(3.46)
0.409
(0.17)
1.443
(1.13)
−0.123
(−0.08)
−7.180
(−1.41)
11.81***
(5.96)
−2.757**
(−2.12)
1.064
(0.51)
0.123
(0.33)
118
0.881
Active developers (per 10,000 inhabitants)
POLS (resid)
OLS
POLS
FD
−0.662
(−0.75)
−1.181
(−0.26)
6.626**
(2.40)
−5.416**
(−2.03)
6.150***
(2.85)
8.155**
(2.44)
0.0669
(0.01)
7.305***
(3.29)
0.942
(0.36)
1.511
(1.08)
0.285
(0.23)
−5.796
(−1.41)
10.01***
(5.97)
−1.929*
(−1.88)
0.920
(0.55)
8.254***
(3.20)
1.686
(0.37)
5.612***
(2.85)
−2.178
(−0.93)
2.116
(1.49)
−0.575
(−0.49)
0.461
(0.07)
5.824*
(1.70)
−1.850
(−0.44)
7.639**
(2.06)
9.126**
(2.33)
−0.740
(−0.12)
7.742***
(2.76)
1.379
(0.42)
1.625
(0.88)
0.871
(1.53)
−0.292
(−1.29)
−0.338
(−0.56)
0.0310
(0.08)
118
0.714
(within)
118
0.607
61
0.850
118
0.884
Registered developers (per 10,000 inhabitants)
POLS (resid)
OLS
POLS
FD
1.630
(0.98)
−3.196
(−0.44)
13.33***
(3.12)
−11.81***
(−2.93)
8.337**
(2.14)
8.105
(1.53)
2.467
(0.32)
5.526
(1.44)
4.358
(0.83)
4.207*
(1.68)
5.869**
(2.46)
−4.148
(−0.46)
19.98***
(6.55)
−5.084**
(−2.57)
1.281
(0.42)
9.821***
(3.12)
1.337
(0.27)
6.885***
(3.28)
−1.346
(−0.56)
2.032
(1.24)
2.024
(0.96)
−1.617
(−0.16)
11.99**
(2.25)
−3.935
(−0.58)
11.90*
(1.88)
9.582
(1.54)
1.245
(0.13)
5.761
(1.24)
5.390
(0.87)
4.544
(1.44)
POLS (resid)
0.731
(1.55)
−0.335
(−1.39)
−0.723
(−0.72)
−0.0177
(−0.03)
−0.211
(−0.21)
−0.236
(−0.36)
118
0.730
(within)
118
0.654
61
0.878
118
0.902
118
0.829
(within)
118
0.227
−0.861
(−0.14)
−4.980
(−0.54)
0.448
(0.11)
7.391
(0.89)
5.803*
(1.82)
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Core developers (per 10,000 inhabitants)
Variation:
Without ‘competition’ and ‘prefer new ideas’
Dependent:
Files
Posts
Core
Active
Registered
Without ‘science adv help’
Files
Posts
Core
Active
Registered
Without ‘IPR protection’
Files
Posts
Core
Active
Registered
regulation
IPR protection
self-det/fulfill
competition
trust
prefer new ideas
science adv help
N/N/N
N/Y/N
Y/Y/Y
–
Y/Y/Y
–
N/N/N
Y/N/N
N/Y/N
Y/Y/N
–
Y/Y/Y
–
N/N/N
N/Y/Y
Y/Y/N
Y/Y/Y
–
Y/Y/Y
–
N/N/N
N/Y/Y
Y/Y/N
Y/Y/Y
–
Y/Y/Y
–
N/N/N
N/Y/Y
Y/Y/N
Y/Y/N
–
N/N/N
–
N/Y/Y
N/N/N
N/Y/N
Y/Y/Y
N/N/N
Y/Y/Y
N/N/N
–
Y/N/N
Y/Y/N
Y/Y/N
N/N/N
Y/Y/Y
N/N/N
–
N/Y/Y
Y/Y/N
Y/Y/Y
N/N/N
Y/Y/Y
N/N/N
–
N/Y/Y
Y/Y/N
Y/Y/Y
N/N/N
Y/Y/Y
N/N/N
–
N/Y/Y
Y/Y/N
N/N/N
N/N/N
N/N/N
N/N/N
–
N/N/N
–
Y/Y/Y
N/N/N
Y/Y/Y
N/N/N
N/N/N
Y/Y/N
–
Y/Y/N
N/N/N
Y/Y/Y
N/N/N
N/N/N
N/Y/Y
–
Y/Y/Y
N/N/N
Y/Y/Y
N/N/N
N/N/N
N/Y/Y
–
Y/Y/Y
N/N/N
Y/Y/Y
N/N/N
N/N/N
N/Y/Y
–
Y/Y/N
N/N/N
N/N/N
N/N/N
N/N/Y
files = uploaded files, posts = posted messages, core = core developers, active = active developers, registered = registered developers. Y = Yes and N = No with scheme: OLS/POLS/FD or residual POLS.
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Table 5
Significance (p < 0.10) of explanatory variables in further model variations.
103
104
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
measure and (b) our principal component ‘self-det/fulfill’. The results of these two variations are very close, indicating that
our ‘self-det/fulfill’ measure indeed taps the same cultural dimension as individualism.25
So our findings support both research emphasizing the importance of intrinsic self-fulfillment motives and the literature
emphasizing the importance of trust. It is interesting to know whether the combination of both is relevant. Therefore, we also
run regressions where we include an interaction term ‘trust and self’. The results provide some evidence that it is indeed the
combination of trust and self-determination/fulfillment that matters: for posted messages, core and active developers the
interaction term is positively significant in the OLS and POLS regressions, while ‘self-det/fulfill’ alone becomes insignificant.
Thus, for these cases, it holds that self-determination/fulfillment alone has no impact, it is only beneficial if it is combined
with trust.
While our hypothesis focusing on intrinsic motives is supported, we have to reject Hypothesis 2, stating that a culture
of positive attitudes toward competition has a positive impact on the number of OSS developers and on the OSS activity
level. The explanatory variable ‘competition’ is never significant. We thus also run regressions without ‘competition’ as well
as regressions without ‘competition’ and ‘prefer new ideas’ (the latter reported in Table 5). The results for the remaining
explanatory variables stay quite robust. The fact that Hypothesis 2 has to be rejected must be interpreted with care. As
mentioned above, finding a positive effect of positive attitudes toward competition may well show the relevance of external
incentives, albeit favoring competition is not a logically necessary condition or precondition. Some individuals who engage
in OSS because of external motives might nevertheless reject the idea that competition in general is good, for example, if
they try to achieve reputation signals for the job market, because they know that they have to compete against very good
job candidates, and they fear that they may otherwise have no chance. Also those developers for whom reputation signals
are just a byproduct of internally motivated contributions may reject the idea of competition. What we can say is that we
find no support for the hypothesis that OSS development is driven by external motives because OSS community members
favor competition (and therefore want to achieve signals). This may explain the mismatch between the ranking of extrinsic
motivations in the self-reporting of developers and its importance when it comes to actual performance.
For the hypothesis on regulation—which is linked to career and commercial motives—we do find some support: we
cannot reject Hypothesis 3 with respect to the number of developers, since ‘regulation’ is negatively correlated in the POLS
and the FD models regarding the number of core developers, active developers and registered users. Regarding OSS activities,
we find effects only in the OLS of posted messages. But as OLS is the weakest of the three estimation methods we use, we
have to treat this result with care. Since we control for Internet access, we use ‘regulation’ as a proxy for the influence of
the ICT sector (and for all the cases where regulation is significant, we also run regressions with an interaction term of
regulation with Internet access: the results are the same). The quality of our measure may be criticized, but in the absence
of a better measure,26 we have to refer to the regulation results. So we do find evidence that the importance of an ICT sector
translates into a positive impact on the number of OSS developers—albeit not on their activities (at least not when it comes
to uploading files, our strongest measure of OSS activities). Does this mean that extrinsic motives are important for the
decision to become an OSS developer, but not for actual activity levels? This would contradict previous research finding the
opposite, since Hars and Ou (2002) find a positive relation between extrinsic motives and effort, although these motives
are typically ranked relatively low in self-reports. We rather attribute our results to the quality of our proxy. Comparing
countries’ economic regulation (in addition to Internet access) certainly does capture some differences in their ICT sectors,
but this is not at all precise enough to capture the differences in the career opportunities the various ICT sectors offer. A much
more precise measure may thus also be significant with respect to activities. Nevertheless, we do measure some differences
between the countries’ ICT sectors, but the positive correlation of low regulation and the number of OSS developers may
exist simply because with a growing ICT sector, more people get involved in software programming in general, and some of
them become OSS developers. Moreover, there is also the impact of OSS business models. ICT firms who use OSS take part
in the OSS community (either directly or indirectly by paying, e.g., core developers) and thus have an impact on the number
of core, active, and registered developers, but their contributions may not be relevant enough to create a significant signal.
The variable ‘IPR protection’ is significant in the POLS regressions of the two activity measures, and in the OLS and POLS
regressions of all developer measures. The reason why it is not significant in the FD regressions may be that the within
variation of ‘IPR protection’ is too low given our time periods. When we treat IPR protection as a time-invariant variable, and
hence, delete it from the FD but include the 2004 values in the FD-residual regressions, IPR becomes significant. Moreover,
the VIF of ‘IPR protection’ is above 10 for the OLS regressions. So we may underestimate the true relationship because of
multicollinearity. This could explain why ‘IPR protection’ is not significant in the OLS regression of the number of posted
files.27 For robustness check, we also run regressions without ‘IPR protection’, the results are basically the same (see Table 5).
25
The measure for post-materialism alone is “weaker”—since its significance is lower than that of ‘self-det/fulfill’. So while the results of the FD-residual
POLS are the same as for ‘self-det/fulfill’, it becomes insignificant for some of the other regressions. But this is not surprising since post-materialism is
only one of the variables that belong to the higher-level concept of self-expression, which is equivalent to individualism (Inglehart, 2006; Inglehart and
Oyserman, 2004). Therefore one would have expected post-materialism alone to be weaker.
26
The only ICT-related measures that are available are (a) ICT service exports and (b) ICT goods exports, both as a share of total service and goods exports.
We decided not to use these measures because they capture the export orientation of a country’s ICT sector rather than the sector itself. Moreover, only
the share of ICT service of total service exports is significant in some regressions.
27
We run regressions without GDP, since this brings the VIF of ‘IPR protection’ close to 10. However, ‘IPR protection’ stays non-significant in the posted
files OLS regression.
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
105
So we cannot reject Hypothesis 5, stating that the protection and enforceability of IPRs has a positive impact on the number
of OSS developers and on the OSS activity level. This is a clear rejection of the view that OSS is anti-IPR since it “frees”
software from intellectual property. In this view, the lack of enforceability of intellectual property should be irrelevant (if
not positive) for OSS. But the opposite holds, and the denial of IPR as such may even harm the supply of OSS. We have to
discuss an often-mentioned objection in this context: The argument is that in societies with a low de facto protection of
IPRs, there is less need for OSS because one can get software for free (or at a low cost) anyway. This will explain why we
have greater OSS contribution when IPR protection is strong. This argument sees OSS as a substitute for pirated software.
However, this explanation is not convincing because OSS is far more than just “cheap” software: the key element of OSS is
that one has access to the source code and can thus further develop it. A pirated copy of proprietary software is still just a
copy of the binary code. Whereas the source code is the human-readable recipe of software, the binary code is not readable
by humans. Thus, pirated software cannot be a substitute for OSS as it is missing the source code. Finally, the objection mixes
up demand and supply side arguments. The argument points to the demand-side, but we analyze the supply side of OSS:
activity levels and the number of developers.
There may be a downside because ‘IPR protection’ measures the enforceability of intellectual property in general, while
different forms of IPR may affect OSS in different ways. For example, software patents are often perceived as being a threat
to the OSS movement (Hall and MacGarvie, 2010), since open source volunteers and small and medium-sized enterprises
can afford neither patent portfolios nor lawsuits (e.g., Vetter, 2009; Välimäki, 2004; Bessen, 2002). In contrast, Fosfuri et al.
(2008) find that firms are more (less) likely to release OSS products when they hold a large stock of software patents
(software trademarks).28 This is because software patents protect the ‘private’ complements that are essential for the firms’
OSS business models, while software trademarks are linked to proprietary business models (Fosfuri et al., 2008). However,
knowing that a society where intellectual property in general is respected is more likely to have OSS activities and developers
nevertheless supports the argument that OSS, in principle, benefits from the general enforceability of intellectual property,
because OSS is a new IPR paradigm and it makes use of IP law. So independent from the fact that different forms of IPR have
different effects on OSS, respect for IPRs in general has a positive effect (at least this statement cannot be rejected).
Finally, with respect to the impact of openness to novelty on OSS, we find that while an openness to new ideas is never
significant, a positive attitude towards scientific progress is significant, but only with respect to the number of registered
developers. So while Hypothesis 6 has to be completely rejected, Hypothesis 7 has to be rejected excepting its relevance
for registered OSS developers. The latter can be interpreted as a sign that there is indeed a need to be open to scientific
progress when it comes to the decision to become an OSS developer (i.e., to enter the OS community)—but when it comes
to differences among the OSS community members (activity levels and roles), openness to scientific progress plays no role.
Our results regarding the openness to novelty and OSS can furthermore lead to the following conclusion. If the claim that
OSS leads to more innovation (Piva et al., 2012; Rossi-Lamastra, 2009; Ghosh, 2006) is right, we find no evidence that this
is due to the fact that OSS attracts more innovative (or open-minded) individuals. This in turn is in line with our general
thrust, as it can be seen as supporting the argument that OSS indeed creates more innovation because of its institutional
arrangements.
To test for robustness we run additional regressions. In Table 5 we report the results of a selection of these variations.
Since neither ‘competition’ nor ‘prefer new ideas’ are significant we excluded both (first model variation reported). We also
run regressions without ‘science adv help’ since this variable is significant only with respect to the number of registered
developers. Finally, as already noted, we report regression results without the measure of enforceability of IPR. In all cases
the results for the remaining explanatory variables stay quite robust.
5. Summary and outlook
This cross-country study shows how the relative number of OSS developers and the OSS activity level of a country depend
on institutional and cultural factors. We connect the microeconomic aspects of OSS—which in Williamson’s terms belong to
Levels 3 and 4—with aspects of the institutional environment (Level 2) and culture (Level 1). We discuss the impact of culture
in terms of values (valuation of competition), preferences (self-determination/fulfillment and openness towards novelty),
and in the form of beliefs (trust), as well as of how culture can affect the functioning of governance institutions (valuation
of competition). In addition, we assess how formal Level 2 institutions have an impact on the lower Levels 3 (enforceability
of IPR) and 4 (regulation as a proxy for a growing ICT industry, which is linked to extrinsic and commercial motives).
A culture characterized by self-determination/fulfillment values is an important positive factor for the supply side of
OSS. This challenges the limited empirical literature linking OSS motives with effort, which claims that self-determination
motives are not important. Our findings are consistent with the results of the self-report surveys, but we are able to avoid
the cultural bias problems these studies may suffer from. Interpersonal trust is another strong explanatory variable. This
supports research results that emphasize the importance of (ex ante general) trust in the trust vs control debate on research
into OSS governance. Moreover, we find some evidence that it is the combination of self-determination/fulfillment and trust
that favors OSS.
28
Harison and Koski (2010) find a similar pattern; however, they use the firms’ records of (general) trademarks and patent applications as an indicator
of intellectual capacity.
106
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
A positive valuation of competition has no significant effect on OSS. So while extrinsic aspects may set incentives (Hars and
Ou, 2002; Hertel et al., 2003), the corresponding motivations may be of a more “reactive” or “passive” kind. To put it simply,
while OSS developers react to external incentives, OSS does not significantly attract individuals who value competition. Here
it would be interesting to see more research on values and the link between extrinsic motives and effort. Moreover, it would
be interesting to see results from cross-country studies that, after correcting for country size, etc., link reported motives
with effort.
We also find some effect of regulation. Since we control for Internet access, we interpret this as a (rough) measure for the
influence of the ICT sector via career motives and commercial aspects. It would be interesting to conduct a study with precise
ICT data for all countries. Further research could involve studying panel data to analyze how changes in the ICT industry
(growth, diffusion of OSS business models, etc.) are correlated with OSS activities.
The positive impact of IPR protection supports the view that OSS is not anti-IPR, but a new IPR paradigm. However,
the enforceability of intellectual property is a broad measure, so it would be fascinating to see studies that link detailed
per-country data on the different IPR laws with OSS. It would also be of interest to analyze whether the enforceability of IPR
in a country has an impact on decisions regarding the type of OSS license project founders from that country adopt.
Finally, openness to novelty does not seem to play an important role for OSS. However, there is a positive impact of
openness towards scientific progress on the number of registered OSS developers. Regarding openness towards new ideas,
we tend to attribute our lack of findings to the quality of our measure. The question asked in the WVS may be too broad
and could be interpreted by the respondents in several ways. It would indeed be much more effective to have measures
of creativity and openness to novelty incorporated in the research on OSS developers. However—as with the motives—the
resulting scores would have had to be corrected for possible sample bias (cultural bias).
Finally, our study not only improves the understanding of the supply side of OSS (i.e., the microeconomics of OSS) but
also contributes to the discussion about OSS as a strategy for developing countries. May (2006), Ghosh (2003) and Weber
(2004b) argue that OSS is a opportunity for developing countries’ ICT sectors, implying that economic actors from developing
countries do not only use OSS, but get actively involved, as developers. Here, institutional and cultural factors could be a
barrier. For example, the actual levels of OSS activity may stay low if there is too little trust. The discussion on strategies to
foster OSS in developing countries (e.g., Yildirim and Ansal, 2011; Câmara and Fonseca, 2007; Li et al., 2004) should take the
importance of institutional and cultural factors into account.
Appendix A.
Table A.1
List of countries (final sample).
No
Country
t=1
t=2
No
Country
t=1
t=2
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Albania
Argentina
Armenia
Australia
Austria
Azerbaijan
Bangladesh
Belgium
Brazil
Bulgaria
Canada
Chile
China
Colombia
Croatia
Czech Republic
Denmark
Dominican Republic
El Salvador
Estonia
Finland
France
Georgia
Germany
Guatemala
Hungary
Iceland
India
Ireland
Italy
Japan
0
1
0
0
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
Latvia
Lithuania
Macedonia
Malta
Mexico
Netherlands
New Zealand
Nigeria
Norway
Pakistan
Peru
Philippines
Poland
Portugal
Romania
Russia
Slovakia
Slovenia
South Africa
South Korea
Spain
Sweden
Switzerland
Turkey
Uganda
Ukraine
United Kingdom
United States
Uruguay
Venezuela
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
107
Table A.2
List of control and explanatory variables.
Variable
Description
Source
log(GDP)
GDP per 100,000 inhabitants
(ppp)
UNDP Education Index
Internet users per 10,000
inhabitants
Index of Regulation of Business
Protection of intellectual
property
World Bank (2007)
self-det/fulfill
Self-determination and
-fulfillment
Own PCA
competition
Attitude towards competition.
WVS (2010), EWVS (2006)
trust
Interpersonal trust
WVS (2010), EWVS (2006)
prefer new ideas
Preference for new ideas
WVS (2010), EWVS (2006)
science adv help
Science advances will help
mankind.
WVS (2010), EWVS (2006)
education
Internet users
regulation
IPR protection
Remark
UNDP (2007, 2005)
ITU (2005, 2003)
Gwartney et al. (2008)
Gwartney et al. (2006)
Inverse version: high scores indicate high regulation
Based on the World Economic Forum Executive Opinion
Survey; indicates whether IPR protection is weak and not
enforced vs is strong and enforced
Principal component based on individualism (Diener et al.,
2000; Oishi, 2000), importance of leisure time,
self-responsibility, children should learn feeling of
responsibility, and Inglehart’s post-materialism (all WVS,
2010; EWVS, 2006).
10-item scale of “Competition is good. It stimulates people to
work hard and develop new ideas” vs “Competition is harmful.
It brings out the worst in people.” Mean value.
Percentage saying “Most people can be trusted” (vs “Can’t be
too careful”)
10-item scale of “Ideas that stood test of time are generally
best” vs “New ideas are generally better than old ones”. Mean
value.
Question: “In the long run, do you think the scientific advances
we are making will help or harm mankind?” Possible answers
are “Will help”, “Will harm”, “Some of each”. Percentage
choosing “Will help”.
Regional dummies skipped.
Table A.3
List of variance inflation factors (VIFs).
OLS
log(GDP)
education
Internet users
regulation
IPR protection
self-det/fulfill
competition
trust
prefer new ideas
science adv help
(regional dummies skipped)
17.26
5.95
9.58
4.31
12.99
5.51
1.63
3.75
1.88
1.85
Observations
61
POLS
13.71
5.06
9.43
3.28
8.82
5.55
1.68
3.67
1.88
1.81
118
POLS (resid)
3.77
1.45
3.30
1.59
1.54
118
Note: Variance inflation factors cannot be computed for first-difference estimations.
References
Abdollahian, M.A., Coan, T.G., Oh, H., Yesilada, B.A., 2012. Dynamics of cultural change: the human development perspective. International Studies Quarterly
56, 827–842.
Alesina, A., Giuliano, P., Nunn, N., 2013. On the origins of gender roles: women and the plough. The Quarterly Journal of Economics 128, 469–530.
Allik, J., Realo, A., 2004. Individualism-collectivism and social capital. Journal of Cross-Cultural Psychology 35, 29–49.
Basabe, N., Ros, M., 2005. Cultural dimensions and social behavior correlates: individualism-collectivism and power distance. Revue Internationale de
Psychologie Sociale/International Review of Social Psychology 18, 189–225.
Bergquist, M., Ljungberg, J., 2001. The power of gifts: organizing social relationships in open source communities. Information Systems Journal 11, 305–320.
Berigan, N., Irwin, K., 2011. Culture, cooperation, and the general welfare. Social Psychology Quarterly 74, 341–360.
Bessen, J., 2002. What good is free software? In: Hahn, R.W. (Ed.), Government Policy Toward Open Source Software. Brookings Institution Press, Washington,
DC.
Bicchieri, C., 2006. The Grammar of Society—The Nature and Dynamics of Social Norms. Cambridge University Press, Cambridge.
Bisin, A., Verdier, T., 2001. The economics of cultural transmission and the dynamics of preferences. Journal of Economic Theory 97, 298–319.
Bitzer, J., Schrettl, W., Schröder, P.J., 2007. Intrinsic motivation in open source software development. Journal of Comparative Economics 35, 160–169.
Bitzer, J., Schröder, P.J.H., 2007. Open source software, competition and innovation. Industry and Innovation 14, 461–476.
Bjørnskov, C., 2007. Determinants of generalized trust: a cross-country comparison. Public Choice 130, 1–21.
Bowles, S., 1998. Endogenous preferences: the cultural consequences of markets and other economic institutions. Journal of Economic Literature 36, 75–111.
Boyce, C.J., Wood, A.M., 2011. Personality and the marginal utility of income: personality interacts with increases in household income to determine life
satisfaction. Journal of Economic Behavior and Organization 78, 183–191.
Breznau, N., Lykes, V.A., Kelley, J., Evans, M.D.R., 2011. A clash of civilizations? preferences for religious political leaders in 86 nations. Journal for the
Scientific Study of Religion 50, 671–691.
108
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Brons, L.L., 2006. Indirect measurement of regional culture in the Netherlands. Tijdschrift voor Economische en Sociale Geografie 97, 547–566.
Câmara, G., Fonseca, F., 2007. Information policies and open source software in developing countries. Journal of the American Society for Information
Science and Technology 58, 121–132.
de Cremer, D., 1999. Trust and fear of exploitation in a public goods dilemma. Current Psychology, 153–163.
Dafermos, G., Sderberg, J., 2009. The hacker movement as a continuation of labour struggle. Capital and Class 33, 53–73.
Dalle, J.M., David, P.A., 2005. Allocation of software development resources in open source production mode. In: Feller, J., Fitzgerald, B., Hissam, S.A., Lakhani,
K. (Eds.), Perspectives on Free and Open Source Software. The MIT Press, Cambridge/London, pp. 297–328.
David, P.A., Rullani, F., 2008. Dynamics of innovation in an ‘open source’ collaboration environment: lurking, laboring, and launching floss projects on
SourceForge. Industrial and Corporate Change 17, 647–710.
De Noni, I., Ganzaroli, A., Orsi, L., 2013. The evolution of OSS governance: a dimensional comparative analysis. Scandinavian Journal of Management 29,
247–263.
Dearmon, J., Grier, K., 2009. Trust and development. Journal of Economic Behavior and Organization 71, 210–220.
Diener, E., Gohm, C.L., Suh, E., Oishi, S., 2000. Similarity of the relations between marital status and subjective well-being across cultures. Journal of
Cross-Cultural Psychology 31, 419–436.
Durkheim, E., 1933. The Division of Labor in Society. Collier-Macmillan, London (Original work published in French 1893).
Economides, N., Katsamakas, E., 2006. Two-sided competition of proprietary vs. open source technology platforms and the implications for the software
industry. Management Science 52, 1057–1071.
Engelhardt, S.v., 2008. Intellectual Property Rights and Ex-Post Transaction Costs: The Case of Open and Closed Source Software. Jena Economic Research
Papers 2008-047. Friedrich-Schiller-University Jena and Max-Planck-Institute of Economics.
Engelhardt, S.v., 2010. Quality Competition or Quality Cooperation? License-Type and the Strategic Nature of Open Source vs. Closed Source Business
Models. Jena Economic Research Papers 2010-034. Friedrich-Schiller-University Jena and Max-Planck-Institute of Economics.
Engelhardt, S.v., Freytag, A., Schulz, C., 2013. On the geographic allocation of open source software activities. International Journal of Innovation in the
Digital Economy 4, 25–39.
EWVS, 2006. European and world values surveys four-wave integrated data file, 1981–2004, v.20060423. Surveys designed and executed by the European
Values Study Group and World Values Survey Association. File Producers: ASEP/JDS, Madrid.
Fehr, E., Schmidt, K.M., 1999. A theory of fairness, competition, and cooperation. The Quarterly Journal of Economics 114, 817–868.
Fernández, R., Fogli, A., 2009. Culture: an empirical investigation of beliefs, work, and fertility. American Economic Journal: Macroeconomics 1,
146–177.
Fershtman, C., Gandal, N., 2008. Microstructure of Collaboration: The ‘Social Network’ of Open Source Software. CEPR Discussion Papers 6789. C.E.P.R.
Discussion Papers.
Fisman, R., Khanna, T., 1999. Is trust a historical residue? Information flows and trust levels. Journal of Economic Behavior and Organization 38, 79–92.
Fitzgerald, B., 2006. The transformation of open source software. MIS Quarterly 30, 587–598.
Fosfuri, A., Giarratana, M.S., Luzzi, A., 2008. The penguin has entered the building: the commercialization of open source software products. Organization
Science 19, 292–305.
Franck, E., Jungwirth, C., 2003. Reconciling rent-seekers and donators-the governance structure of open source. Journal of Management and Governance 7,
401–421.
Furrer, O., Liu, B.S.C., Sudharshan, D., 2000. The relationships between culture and service quality perceptions: basis for cross-cultural market segmentation
and resource allocation. Journal of Service Research 2, 355–371.
Gallivan, M.J., 2001. Striking a balance between trust and control in a virtual organization: a content analysis of open source software case studies.
Information Systems Journal 11, 277–304.
Gambardella, A., Hall, B.H., 2006. Proprietary versus public domain licensing of software and research products. Research Policy 35, 875–892.
Ghosh, R., 2003. Licence Fees and GDP per Capita: The Case for Open Source in Developing Countries. First Monday 8. http://firstmonday.org/
Ghosh, R.A., 2006. Economic impact of open source software on innovation and the competitiveness of the Information and Communication Technologies
(ICT) sector in the EU. Study. European Commission.
Ghosh, R.A., Glott, R., Krieger, B., Robles, G., 2002. FLOSS Final Report, Part 4: Survey of Developers. Technical Report. International Institute of Infonomics,
University of Maastricht.
Giuri, P., Ploner, M., Rullani, F., Torrisi, S., 2010. Skills, division of labor and performance in collective inventions: evidence from open source software.
International Journal of Industrial Organization 28, 54–68.
Giuri, P., Rocchetti, G., Torrisi, S., 2002. Open Source Software: From Open Science to New Marketing Models. LEM Papers Series 2002/23. Laboratory of
Economics and Management (LEM), Sant’Anna School of Advanced Studies, Pisa, Italy.
Gonzalez-Barahona, J.M., Robles, G., Andradas-Izquierdo, R., Ghosh, R.A., 2008. Geographic origin of Libre software developers. Information Economics and
Policy 20, 356–363.
Greif, A., 1994. Cultural beliefs and the organization of society: a historical and theoretical reflection on collectivist and individualist societies. Journal of
Political Economy 102, 912–950.
Guiso, L., Sapienza, P., Zingales, L., 2008. Social capital as good culture. Journal of the European Economic Association 6, 295–320.
Gwartney, J., Lawson, R., Norton, S., 2006. Economic Freedom of the World—2008 Annual Report. The Fraser Institute, Data retrieved from
http://www.freetheworld.com/
Gwartney, J., Lawson, R., Norton, S., 2008. Economic Freedom of the World—2008 Annual Report. The Fraser Institute, Data retrieved from
http://www.freetheworld.com/
Hall, B.H., MacGarvie, M., 2010. The private value of software patents. Research Policy 39, 994–1009.
Harison, E., Koski, H., 2010. Applying open innovation in business strategies: evidence from Finnish software firms. Research Policy 39, 351–359.
Hars, A., Ou, S., 2002. Working for free?—Motivations of participating in open source projects. International Journal of Electronic Commerce 6, 25–39.
Henkel, J., 2006. Selective revealing in open innovation processes: the case of embedded Linux. Research Policy 35, 953–969.
Hertel, G., Niedner, S., Herrmann, S., 2003. Motivation of software developers in open source projects: an Internet-based survey of contributors to the Linux
kernel. Research Policy 32, 1159–1177.
Hippel, E.v., Krogh, G.v., 2003. Open source software and the ‘private-collective’ innovation model: issues for organization science. Organization Science
14, 209–223.
Hofstede, G., 2001. Culture’s Consequences: Comparing Values, Behaviors, Institutions and Organizations Across Nations, 2nd ed. Sage Publications.
Hofstede, G.H., 1991. Cultures and Organizations: Software of the Mind. McGraw-Hill, London/New York.
Inglehart, R., 1977. The Silent Revolution—Changing Values and Political Styles Among Western Publics. Princeton University Press, Princeton.
Inglehart, R., 1981. Post-materialism in an environment of insecurity. The American Political Science Review 75, 880–900.
Inglehart, R., 1990. Culture Shift in Advanced Industrial Society. Princeton University Press, Kassel.
Inglehart, R., 2000. Globalization and postmodern values. The Washington Quarterly 23, 215–228.
Inglehart, R., 2006. Mapping global values. Comparative Sociology 5, 115–136.
Inglehart, R., Baker, W.E., 2000. Modernization, cultural change, and the persistence of traditional values. American Sociological Review 65, 19–51.
Inglehart, R., Foa, R., Peterson, C., Welzel, C., 2008. Development, freedom, and rising happiness: a global perspective (1981–2007). Perspectives on
Psychological Science 3, 264–285.
Inglehart, R., Klingemann, H.D., 2003. Genes, culture, democracy, and happiness. In: Diener, E., Suh, E.M. (Eds.), Culture and Subjective Well-Being. MIT
Press, Cambridge, MA, pp. 165–183.
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
109
Inglehart, R., Oyserman, D., 2004. Individualism, autonomy, self-expression and human development. In: Vinken, H., Soeters, J., Ester, P. (Eds.), Comparing
Cultures: Dimensions of Culture in a Comparative Perspective. Brill, Leiden, pp. 74–96.
Inglehart, R., Welzel, C., 2010. Changing mass priorities: the link between modernization and democracy. Perspectives on Politics 8, 551–567.
Inglehart, R.F., 2008. Changing values among western publics from 1970 to 2006. West European Politics 31, 130–146.
ITU, 2003. International Telecommunication Union ICT Statistics 2003, Part 4—Internet Indicators: Subscribers, Users and Broadband Subscribers. ICT
Statistics. ITU http://www.itu.int/ITU-D/ICTEYE/Reports.aspx
ITU, 2005. International Telecommunication Union ICT Statistics 2005. Part 4—Internet Indicators: Subscribers, Users and Broadband Subscribers. ICT
Statistics. ITU http://www.itu.int/ITU-D/ICTEYE/Reports.aspx
Johnson, J.P., 2002. Open source software: private provision of a public good. Journal of Economics and Management Strategy 11, 637–662.
Johnson, N.D., Mislin, A., 2012. How much should we trust the world values survey trust question? Economics Letters 116, 210–212.
Wendel de Joode, R.V., Bruijn, J.A.d., Eeten, M.J.G.V., 2003. Protecting the Virtual Commons—Self-Organizing Open Source and Free Software Communities
and Innovative Intellectual Property Regimes. T.M.C. Asser Press, The Hague.
Kemmelmeier, M., Jambor, E.E., Letner, J., 2006. Individualism and good works: cultural variation in giving and volunteering across the United States. Journal
of Cross-Cultural Psychology 37, 327–344.
Kimppa, K., 2008. A no-IPR model as solution to reuse and understanding of information systems. In: Avgerou, C., Smith, M., Besselaar, P. (Eds.), Social
Dimensions of Information and Communication Technology Policy. Springer US, pp. 319–325, vol. 282 of IFIP International Federation for Information
Processing.
Kugler, P., 2005. Coordinating Innovation: Evidence from Open Source Software Development. Difo-Druck GmbH.
Kumar, S., 2006. Enforcing the GNU GPL. Journal of Law, Technology and Policy 1, 1–36.
Laat, P.B.d., 2005. Copyright or copyleft? An analysis of property regimes for software development. Research Policy 34, 1511–1532.
Laat, P.B.d., 2007. Governance of open source software: state of the art. Journal of Management and Governance 11, 165–177.
Laat, P.B.d., 2010. How can contributors to open-source communities be trusted? On the assumption, inference, and substitution of trust. Ethics and
Information Technology 12, 327–341.
Lakhani, K.R., Wolf, B., Bates, J., DiBona, C., 2002. The Boston Consulting Group/OSDN Hacker Survey. http://www.osdn.com/bcg/
Lakhani, K.R., Wolf, R.G., 2005. Why hackers do what they do: understanding motivation and effort in free/open source software projects. In: Feller, J.,
Fitzgerald, B., Hissam, S.A., Lakhani, K.R. (Eds.), Perspectives on Free and Open Source Software. MIT Press, pp. 3–22.
Langlois, R.N., Garzarelli, G., 2008. Of hackers and hairdressers: modularity and the organizational economics of open-source collaboration. Industry and
Innovation 15, 125–143.
Lerner, J., Pathak, P.A., Tirole, J., 2006. The dynamics of open-source contributors. American Economic Review 96, 114.
Lerner, J., Tirole, J., 2002. Some simple economics of open source. Journal of Industrial Economics 50, 197–234.
Lerner, J., Tirole, J., 2005. The scope of open source licensing. Journal of Law, Economics and Organization 21, 20–56.
Lewis, J.A., 2008. Intellectual Property Protection—Promoting Innovation in a Global Information Economy. A Report of the CSIS Technology and Public
Policy Program, Centre for Strategic and International Studies. CSIS Press, Washington, DC.
Li, M., Lin, Z., Xia, M., 2004. Leveraging the open source software movement for development of china’s software industry. Information Technologies and
International Development 2, 45–64.
Li, T., Fung, H.H., 2013. Age differences in trust: an investigation across 38 countries. The Journals of Gerontology Series B: Psychological Sciences and Social
Sciences 68, 347–355.
Lindberg, V., 2008. Intellectual Property and Open Source—A Practical Guide to Protecting Code. O’Reilly.
Llanes, G., de Elejalde, R., 2013. Industry equilibrium with open-source and proprietary firms. International Journal of Industrial Organization 31, 36–49.
Marini, M., 2013. The traditions of modernity. The Journal of Socio-Economics (in press).
Markus, M.L., 2007. The governance of free/open source software projects: monolithic, multidimensional, or configurational? Journal of Management and
Governance 11, 151–163.
Maurer, S.M., Scotchmer, S., 2006. Open source software: the new intellectual property paradigm. In: Hendershott, T. (Ed.), Handbook of Economics and
Information Systems. Elsevier, pp. 285–319.
May, C., 2006. The FLOSS alternative: TRIPs, non-proprietary software and development. Knowledge, Technology and Policy 18, 142–163.
McGowan, D., 2001. The legal implications of open source software. Illinois Law Review, 241–304.
Mustonen, M., 2003. Copyleft-the economics of Linux and other open source software. Information Economics and Policy 15, 99–121.
Nevitte, N., Kanji, M., 2002. Authority orientations and political support: a cross-national analysis of satisfaction with governments and democracy.
Comparative Sociology 1, 387–412.
Nicolás, J.D., 1996. Social position, information and postmaterialism. Revista Espa nola de Investigaciones Sociológicas (English Edition of 1996), 153–165.
Norris, P., Inglehart, R., 2002. Islamic culture and democracy: testing the ‘clash of civilizations’ thesis. Comparative Sociology 1, 235–263.
Oishi, S., 2000. Goals as cornerstones of subjective well-being: linking individuals with cultures. In: Diener, E., Suh, E. (Eds.), Cross-Cultural Psychology of
Subjective Well-Being. MIT Press, pp. 87–112.
O’Mahony, S., 2003. Guarding the commons: how community managed software projects protect their work. Research Policy 32, 1179–1198.
O’Mahony, S., Ferraro, F., 2007. The emergence of governance in an open source community. Academy of Management Journal 50, 1079–1106.
O’Mahony, S., Ferraro, F., 2012. Managing the boundary of an ‘open’ project. In: Padgett, J., Powell, W. (Eds.), The Emergence of Organizations and Markets.
Princeton University Press, Princeton and Oxford, pp. 545–565.
Osterloh, M., Rota, S., 2004. Trust and community in open source software production. Analyse and Kritik. Zeitschrift für Sozialtheorie 26, 279–301.
Ostrom, E., 1998. A behavioral approach to the rational choice theory of collective action. American Political Science Review, 1–22.
Paun, F., Tunzelmann, N.v., Richard, P., 2012. Asymmetries and dynamic interactive capabilities in technology transfer between ONERA—the French
aerospace lab and SMEs. Journal of Innovation Economics 9, 103–137.
Perkins, G., 1998. Open Source and Capitalism, Online article posted at slashdot.org: http://slashdot.org/articles/980824/0854256.shtml
Pisano, G., 2006. Profiting from innovation and the intellectual property revolution. Research Policy 35, 1122–1130.
Piva, E., Rentocchini, F., Rossi-Lamastra, C., 2012. Is open source software about innovation? Collaborations with the open source community and innovation
performance of software entrepreneurial ventures. Journal of Small Business Management 50, 340–364.
Polanski, A., 2007. Is the General Public Licence a rational choice? Journal of Industrial Economics 55, 691–714.
Putnam, R.D., 1993. The prosperous community: social capital and public life. American Prospect 4, 35–42.
Putnam, R.D., 1995. Bowling alone: America’s declining social capital. Journal of Democracy 6, 65–78.
Rabin, M., 1993. Incorporating fairness into game theory and economics. American Economic Review 83, 1281–1302.
Ramanujam, P., 2007. Who Produces Open Source Software. MPRA Paper 4253. University Library of Munich, Germany.
Realo, A., Allik, J., Greenfield, B., 2008. Radius of trust: social capital in relation to familism and institutional collectivism. Journal of Cross-Cultural Psychology
39, 447–462.
Realo, A., Koido, K., Ceulemans, E., Allik, J., 2002. Three components of individualism. European Journal of Personality 16, 163–184.
Reisinger, M., Ressner, L., Schmidtke, R., Thomes, T.P., 2013. Crowding-In of Complementary Contributions to Public Goods. Working Paper. WHU—Otto
Beisheim School of Management.
Rossi-Lamastra, C., 2009. Software innovativeness. A comparison between proprietary and free/open source solutions offered by Italian SMEs. R&D Management 39, 153–169.
Sandholtz, W., Taagepera, R., 2005. Corruption, culture, and communism. International Review of Sociology 15, 109–131.
Sen, A.K., 1974. Choice, orderings and morality. In: Krne, S. (Ed.), Practical Reason. Yale University Press, New Haven, London.
110
S.v. Engelhardt, A. Freytag / Journal of Economic Behavior & Organization 95 (2013) 90–110
Sen, R., Subramaniam, C., Nelson, M.L., 2008. Determinants of the choice of open source software license. Journal of Management Information Systems 25,
207–239.
Sharma, S., Sugumaran, V., Rajagopalan, B., 2002. A framework for creating hybrid-open source software communities. Information Systems Journal 12,
7–25.
Shulruf, B., Hattie, J., Dixon, R., 2007. Development of a new measurement tool for individualism and collectivism. Journal of Psychoeducational Assessment
25, 385–401.
Sinha, V.S., Mani, S., Sinha, S., 2011. Entering the circle of trust: developer initiation as committers in open-source projects. In: Proceedings of the 8th
Working Conference on Mining Software Repositories, ACM, New York, NY, pp. 133–142.
Stewart, K.J., Gosain, S., 2006. The impact of ideology on effectiveness in open source software development teams. MIS Quarterly 30, 291–314.
Suh, E., Diener, E., Oishi, S., Triandis, H.C., 1998. The shifting basis of life satisfaction judgments across cultures: emotions versus norms. Journal of Personality
and Social Psychology 74, 482–493.
Tabellini, G., 2008. Institutions and culture. Journal of the European Economic Association 6, 255–294.
Tabellini, G., 2010. Culture and Institutions: Economic Development in the Regions of Europe. Technical Report 4.
Triandis, H.C., 1995. Individualism and Collectivism. Westview Press, New York.
Triandis, H.C., Gelfand, M.J., 2012. A theory of individualism and collectivism. In: Paul, A.M., Van Lange, Arie, W., Kruglanski, E.T.H. (Eds.), Handbook of
Theories of Social Psychology, vol. 2. Sage, Los Angeles, London, Washington, DC, pp. 498–520.
Ullmann-Margalit, E., 1977. The Emergence of Norms. Oxford University Press, Oxford.
UNDP, 2005. Human Development Report 2005. United Nations Development Programme. Palgrave Macmillan.
UNDP, 2007. Human Development Report 2007/08. United Nations Development Programme. Palgrave Macmillan.
Välimäki, M., 2003. Dual licensing in open source software industry. Systemes d’Information et Management 8, 63–75.
Välimäki, M., 2004. A practical approach to the problem of open source and software patents. European Intellectual Property Review 26, 523–527.
Vetter, G.R., 2009. Commercial free and open source software: knowledge production, hybrid appropriability, and patents. Fordham Law Review 77,
2087–2141.
Weber, K., 2004a. Social aspects of non-proprietary software. International Review of Information Ethics, 2.
Weber, S., 2004b. The Success of Open Source. Harvard University Press.
Welzel, C., 2010. How selfish are self-expression values? A civicness test. Journal of Cross-Cultural Psychology 41, 152–174.
Welzel, C., Inglehart, R., Deutsch, F., 2005. Social capital, voluntary associations and collective action: which aspects of social capital have the greatest ‘civic’
payoff? Journal of Civil Society 1, 121–146.
Williamson, O.E., 2000. The new institutional economics: taking stock, looking ahead. Journal of Economic Literature 38, 595–613.
World Bank, 2007. International Comparison Program Database: World Development Indicators.
WVS, 2010. World Values Survey 1981–2008 Official Aggregate v. 20090901. World Values Survey Association www.worldvaluessurvey.org
Yamagishi, T., Cook, K.S., Watabe, M., 1998. Uncertainty, trust, and commitment formation in the United States and Japan. American Journal of Sociology
104, 165–194.
Yamagishi, T., Kanazawa, S., Mashima, R., Terai, S., 2005. Separating trust from cooperation in a dynamic relationship: prisoner’s dilemma with variable
dependence. Rationality and Society 17, 275–308.
Yamagishi, T., Yamagishi, M., 1994. Trust and commitment in the United States and Japan. Motivation and Emotion 18, 129–166.
Yildirim, N., Ansal, H., 2011. Foresighting FLOSS (Free/Libre/Open Source software) from a developing country perspective: the case of turkey. Technovation
31, 666–678.