Outward foreign direct investment and domestic, regional
innovation performance in China: the role of OFDI country
destination diversity and region-specific institutions
Han-Ying Lee 11371609
June 23, 2017, Final
MSc. in Business Administration - International Management Track
ABS, UvA
Supervisor: Dr. Mashiho Mihalache
Statement of originality
This document is written by Han-Ying Lee who declares to take full responsibility for
the contents of this document.
I declare that the text and the work presented in this document is original and that no
sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of
Table of contents
Abstract ... 4
1. Introduction ... 5
2. Literature review ... 11
2.1. Outward foreign direct investment ... 11
2.2. FDI spillovers and OFDI reverse spillovers ... 14
2.3. The effects of institutions on foreign direct investment ... 18
2.4. Regional innovation system in emerging markets ... 21
3. Conceptual framework and hypotheses ... 26
3.1. OFDI country destination diversity and domestic, regional innovation ... 26
3.2. Moderating effect of intellectual property right protection ... 30
3.3. Moderating effect of market development ... 33
3.4. Moderating effect of international openness ... 36
4. Methodology ... 38
4.1. Sample and data collection ... 38
4.2. Model specification ... 40
4.3. Estimation methodology ... 42
4.4. Variables and measures ... 44
4.4.1. Dependent variable ... 44
4.4.2. Independent and moderating variables ... 44
4.4.3. Control variables ... 47
5. Findings ... 50
6. Discussion ... 59
6.1. Theoretical implications ... 61
6.2. Implications for practice ... 64
6.3. Limitations and future research ... 65
7. Conclusion ... 67
Abstract
Recent years outward foreign direct investment (OFDI) flows from emerging economies
have increased substantially, but do these overseas investments bring innovation benefits at
home? In this paper, we aim to shed new light on this phenomenon and contribute to
disentangling the innovation performance implications of OFDI reverse spillovers. We
advance prior foreign direct investment (FDI) spillovers literature by examining the effect of
the diversity of OFDI country destinations on domestic, regional innovation performance,
and how institutions at sub-national level moderate this relationship. Using panel data of
OFDI activities from 30 Chinese provinces during the period 2011-2014, we find that the
diversity of OFDI country destinations is positively related to domestic, regional innovation
performance. We also find that this positive relationship is stronger in regions with higher
levels of intellectual property right enforcement, while weaker in regions with higher levels
of international openness.
Keywords: OFDI reverse spillovers, the diversity of OFDI country destinations,
1. Introduction
According to 2016 World Investment Report conducted by UNCTAD, outward foreign
direct investment (OFDI) flows from emerging economies have increased considerably from
$91 billion to $400 billion from 2000 to 2015, and now account for one-third of global
foreign direct investment (FDI) flows. Among which, China invested more OFDI than any
other emerging economy and was the third largest source of FDI outflows in the world by the
end of 2015 (UNCTAD, 2016).
One major motivation of OFDI by emerging economy multinational enterprises
(EMNEs) is to access strategic assets such as superior technologies in developed countries
and transfer the technologies back to the home country (Luo & Tung, 2007). These positive
externalities that benefit domestic firms with the presence of FDI are defined as "spillovers"
(Blomström, 1986). EMNEs often expect a broader spillovers effect of OFDI on domestic
firms, that is, reverse spillover (Li, Li, Lyles, & Liu, 2016), which can result in productivity
increase among domestic firms (Xia, Ma, Lu, & Yiu, 2014). The impact of FDI on domestic
firms’ productivity has drawn increasing research interest over the past decade (e.g. Haskel,
However, Zhang, Li, Li, and Zhou (2010) suggest that future research should examine
the effect of FDI spillovers beyond domestic firms’ productivity; for instance, exploring how
FDI can influence domestic firms' innovation will be interesting. Some studies have started to
address this call; nevertheless, in contrast to the large amount of research on the effects of
inward FDI (IFDI) on the innovation performance of host economies (e.g. García, Jin, &
Salomon, 2013), very few studies have examined the impact of OFDI on the innovation
performance of home economies in the context of emerging economies (e.g. Deng, 2007).
What's more, despite the appealing argument that FDI generates spillovers that benefit
domestic firms, previous studies have produced mixed results on this topic in emerging
markets, including positive effects (Wang, Deng, Kafouros, & Chen, 2012), marginal effects
(Haskel et al., 2007), and even negative effects (Jordaan, 2008).
There are three limitations of existing studies that can help to explain the underlying
reasons for the mixed results. First, Görg and Strobl (2001) argue that most previous studies
on FDI spillovers mainly focus on the simpler issue of whether the presence of FDI has an
impact on firms’ productivity. These studies have considered FDI as homogenous flows of
capital and have mostly overlooked the heterogeneous nature of FDI, such as entry modes of
Second, existing research on China has overlooked the link between institutions and FDI
spillovers. Institutions, defined as “the rules of the game” (North, 1990: 3) shaping firm
behavior, interaction, and performance, consequently, are crucial factors that influence FDI
spillovers that benefit domestic firms. Since both foreign and local firms have to adapt to
institutional conditions that vary among countries, it is important to consider the role of
institutions when examining FDI spillovers in emerging economies, where institutions are
complicated and changing rapidly (Yi, Chen, Wang, & Kafouros, 2015). However, prior
research mainly focuses on aggregate institutions at country-level, assuming that institutions
are homogenous across subnational regions in a country. While for countries with
heterogeneous institutions across subnational regions, such as China, the subnational effects
are critical (Yi et al., 2015). Existing literature provides limited knowledge of how such
cross-regional institutional variations affect FDI spillovers on domestic firms.
Third, the extant literature on FDI spillovers effects on productivity mainly focuses on
multiple sectors or countries level. However, when examining innovation performance, there
is a considerable literature proposing that innovation performance varies not just between
countries, but also between subnational regions (e.g. Fritsch, 2002). The reasons underlying
Regional innovation system is a system in which firms are systematically engaged in
interactive learning through an embedded institutional environment (Cooke, Uranga, &
Etxebarria, 1998). RIS theory is particularly suitable when studying the determinants of
innovation performance in the context of countries that contain substantial regional
disparities. For example, China, which is diverse in geography, economy, and innovative
capabilities across provinces (Fu, 2008).
Hence, this study aims to address these three gaps. By going beyond the existing
literature, our research question is to explore how the portfolio of OFDI can affect domestic
innovation and how institutions can influence this relationship. Specifically, we investigate
the effect of the diversity of OFDI country destinations on domestic, regional innovation
performance, and how region-specific institutions affect this relationship. We define the
diversity of OFDI country destinations as the extent to which OFDI activities of a province
are formed in different host countries. The institutional constructs contain different
dimensions of region-specific economic institutions, including intellectual property right
protection, market development, and international openness. These different factor
subsidiaries to parent firms but also the absorptive capacity of parent firms and thus the OFDI
spillovers effects.
To test the impact of OFDI country destination diversity on domestic, regional
innovation performance and the institutional factors that moderate this relationship, we
conduct Generalized Estimating Equations (GEE) analysis on a panel data of OFDI activities
from 30 Chinese provinces from 2011 to 2014. China is a particularly appropriate context for
this study not only because the country is the third largest sources of OFDI in the world but
also because it has gone through significant political and economic reforms that have resulted
in varying provincial institutions (Hong, Wang, & Kafouros, 2014). Our results indicate that
the heterogeneous nature of OFDI can have a different impact on OFDI reverse spillovers
and that home region-specific institutions can affect the occurrence and magnitude of this
reverse knowledge spillovers.
This study has multiple contributions to the existing literature. First, we shift the focus
from IFDI to OFDI, from developed economies to emerging economies, given that the FDI
literature has yet to examine the potential reverse spillover benefits of OFDI on the EMNE parent firms (Globerman & Chen, 2010; Meyer & Sinani, 2009). The results provide
positive effects of the diversity of OFDI country destinations on domestic, regional
innovation performance. To the best of our knowledge, this is the first study that examines
the effects of the heterogeneous nature of OFDI concerning country diversity rather than the
simple presence of OFDI on domestic, regional innovation. Second, it is among the first
attempts to theorize the roles played by home region-specific institutions in OFDI reverse
spillover effects. We demonstrate the vital role of institutions in facilitating the absorption of
the knowledge derived from Chinese OFDI. In particular, our study examines how regional
specific institutions facilitate or constrain the extent to which domestic regions benefit from
OFDI. Third, by modeling spillovers in the province-level, we also illuminate how the effects
of OFDI vary across subnational regions. For practical implications, our findings suggest that
policymakers in emerging markets need to develop policies that promote not only a large
amount of OFDI but also OFDI in diverse countries to improve domestic innovation. At the
same time, it is also important for policymakers to consider institutional conditions to enlarge
the benefits of reverse spillovers from OFDI. Overall, this study complements existing
research on FDI spillovers and sheds new light on the innovation performance implications of
The remainder of the paper is structured as follows. Section 2 reviews the relevant
literature on OFDI, FDI spillovers, the effects of institutions on FDI, and regional innovation
system in emerging economies, and poses the research question. Section 3 introduces a
conceptual framework of how the diversity of foreign investments interacts with
location-bound institutions to jointly shape the effects of OFDI reverse spillovers on
domestic, regional innovation performance. We then present our methodological approach
and estimating methodology in section 4. Section 5 presents the results of our analysis. In
section 6, we discuss the theoretical and managerial implications of our findings, and the
main limitations of our study with several suggestions for future research. Finally, section 6
forms a conclusion and highlights the contributions of this study.
2. Literature review
2.1. Outward foreign direct investment
Foreign direct investment (FDI) is defined as an investment made to acquire lasting
interest in enterprises operating outside of the economy of the investor (Moran, 2012). In
international business literature, it is customary to classify the motivation of OFDI as natural resource seeking, market seeking, efficiency seeking, or strategic asset seeking
of the most important motivations for EMNEs to overcome their latecomer disadvantages,
and this is more particularly for Chinese MNEs (Child & Rodrigues, 2005; Luo & Tung,
2007). Rui and Yip (2008) argue that Chinese firms often use cross-border acquisitions to
gain strategic assets to compensate for their competitive disadvantages, while simultaneously
leveraging their unique ownership advantages.
Past research has revealed that EMNEs acquire technology from developed countries to
enhance their competitiveness back in their home markets also in the international market (Li,
Li, & Shapiro, 2012; Peng, 2012). Since technological resources are concentrated in
developed countries (Keller, 2004), investing in these countries and locating proximate to
technology centers can increase the opportunities for exchange and learning of localized tacit
knowledge and technical (Fosfuri, Motta, & Rønde, 2001).
There is empirical evidence on the relationship between OFDI by developed market
MNEs and their home country productivity growth, while the results on productivity growth
are not consistent. For example, De la Potterie and Lichtenberg (2001) find that both imports
and OFDI contribute to productivity growth in the United States, Japan, and 11 European
countries over the period 1971-1990. Driffield, Love, and Taylor (2009) distinguish between
types of outward investments increase productivity growth back in the United Kingdom. In
contrast, by examining Swedish MNEs' OFDI both in a firm- and industry-level, Braconier,
Ekholm, and Knarvik (2001) find no correlation between OFDI and their domestic
productivity. Similarly, Bitzer and Görg (2009) study 17 OECD countries' OFDI reverse
spillovers by using industry-level data. They detect that OFDI has a negative effect on some
of the OECD countries productivity, such as Denmark, Italy, South Korea, Norway, and
Spain. The OFDI effects are widely different across countries.
While mixed findings on the relationship between OFDI and home country productivity
growth are revealed in the context of developed countries, few studies test such relationships
in emerging countries, with only a few exceptions. Zhao, Liu, and Zhao (2010) examine the
contribution of Chinese OFDI to domestic productivity changes. They confirm the beneficial
spillover effects of OFDI on home country productivity growth over the period from 1991 to
2007, and that gains in efficiency have been the chief reason for this. In a similar vein, Herzer
(2011) investigate the relationship between OFDI and home country productivity by using a
sample of 33 developing countries for the period 1980 to 2005. They find that OFDI had a
robust and positive long-term effect on domestic productivity and that different extent of
and Görg’s (2009) study, they find that in South Korea, OFDI has the largest negative effect
on its domestic productivity.
Given the mixed results of the effects of OFDI on domestic productivity, Görg and
Strobl (2001) explain the underlying reasons. They argue that most previous FDI spillovers
research mainly "focus on the simpler issue of whether the presence of FDI affects
productivity in domestic firms" (Görg & Strobl, 2001: 724). These studies have considered
FDI as homogenous flows of capital and have mostly overlooked the heterogeneous nature of
FDI, such as entry modes of firms and the origins of countries (Zhang et al., 2010). Therefore,
in this study, we go beyond the existing literature, which mainly focuses on the FDI presence
in the industry, and further examine how the portfolio of OFDI influences innovation at home
market.
2.2. FDI spillovers and OFDI reverse spillovers
The mechanism underlying the relationship between FDI and domestic productivity is
knowledge spillover effects (Zhang et al., 2010). FDI spillovers refer to positive externalities
that benefit domestic firms with the presence of FDI (Zhang et al., 2010). When MNCs enter
a host country, they deliver capital inflows, raise local employment and bring technological
The existing literature on IFDI recognizes four main channels for spillover effects to
occur. First, demonstration effect, through which domestic firms can improve their
productivity by observing and imitating the technology and management practices of foreign
firms (Kokko, 1994). Second, employment turnover, when employees from foreign affiliates
move to domestic firms, the technologies and management practices of foreign firms can
diffuse to domestic firms (Kokko, Tansini, & Zejan, 1996). Third, industry linkage, when
foreign affiliates build backward and forward linkages with domestic suppliers and
distributors, knowledge can be transmitted to domestic firms using the same suppliers and
distributors (Spencer, 2008). The fourth channel is through competition. The entry of foreign
affiliates can force domestic firms to increase their productivity by updating technologies and
adopting management practices to respond to competitive challenges (Blomström & Kokko,
1998). While the effects of competition can also be negative, as the entry of foreign may steal
the markets of final goods or raw materials away from their domestic competitors, this is
called a crowding out effect (Aitken & Harrison, 1999).
Most conventional research of knowledge spillovers focuses on the effects of IFDI on
host country in emerging economies (Li, Zhang, & Lyles, 2013). While the significance of
because this phenomenon is rather new and it is hard to obtain reliable empirical data (Lyles,
Li, & Yan, 2014).
Nevertheless, a group of studies starts to shift the light of spillover effects from IFDI to
OFDI (e.g. De la Potterie & Lichtenberg, 2001; Driffield et al., 2009; Herzer, 2011). Scholars
argue that the logic of spillover effects from inward and outward FDI is similar and that those
indicated channels in FDI spillovers above are also influential for OFDI (Chen, Li, & Shapiro,
2012). First, when EMNEs invest in developed markets, demonstration effects can be
realized through learning and imitating advanced technological know-how and management
techniques from local firms in developed economies (Herzer, 2011). Second, subsidiaries
abroad will employ local technical and managerial staffs, which will improve subsidiaries
technology and managerial practices, in turn, the parent firm (Singh & Agrawal, 2011).
Chinese MNEs for instance, they set up R&D centers in developed markets to enhance their
embeddedness in local innovation systems to learn (Di Minin, Zhang, & Gammeltoft, 2012).
Third, subsidiaries may acquire technical guidance and support through backward and
forward linkages with host country firms. For example, Harhoff, Mueller, and Reenen (2014)
show that German subsidiaries in the United States (US) benefit from R&D knowledge
downstream firms helps German firms to better tap into the R&D resource. Fourth,
competition will force EMNEs to update their technologies and practices to become more
productive because they face more competitive pressures in developed economies than in
their home economies (Li et al., 2016).
As a result, MNEs can transfer strategic assets accessed in oversea markets back to
home country, and so MNE parent firms can benefit from OFDI activities through these
channels, this is called reverse spillover (Li et al., 2016). Positive reverse spillover effects are
realized through knowledge spillovers to EMNE subsidiaries in a developed economy and
subsequently transferred from these subsidiaries back to their parent firms in emerging
economies (Chen et al., 2012).
However, past empirical results do not always confirm the existence of spillovers theory.
Discovering that FDI spillover effects are not consistent as many studies assume, more recent
research has examined the conditions under which spillovers might be strong, weak or
nonexistent. For example, Li et al. (2016) find that reverse knowledge spillovers will only
occur when the technological gap between a home province and MNEs' host countries is not
too large; otherwise, MNE parent firms are not able to capture the benefits of reverse
reverse spillovers are enhanced in home provinces with a greater presence of IFDI thanks to
the beneficial demonstration effects.
These findings from previous studies imply that location-bound factors could influence
the extent to which parent firms benefit from OFDI. Nevertheless, despite the concept that
differences in geography can explain variations in the effects of IFDI (Yi et al., 2015), prior
literature has not studied how such region-specific institutions affect OFDI reverse spillovers.
In this study, we investigate the effect of regional institutions on the occurrence and
magnitude of OFDI reverse spillovers across regions by exploring how they influence parent
firms as spillover receivers.
2.3. The effects of institutions on foreign direct investment
North (1990) defines institutions as the set of fundamental political, social, and legal
ground rules in the country, including both formal and informal constraints that shape human
interaction in society. The institution-based view assumes that firm-level actions are a
function of pressures in the firm's external environment. This view underlines the
complicated and changing relationships between organizations and their surrounding
institution-based view has become one of the leading viewpoints for theorizing international
strategy such as why and how firms conduct FDI (Peng, Wang, & Jiang, 2008).
The application of the institution-based view in studying FDI issues is dominated by two
groups of literature. The first group focuses on host-region institutions. As institutions create
both opportunities and barriers for IFDI, they constitute a critical portion of locational
advantages (Dunning, 1998). This literature suggests that more efficient institutions reduce
transaction costs and uncertainties of doing business in the host region, and thus providing
incentives for IFDI and entry by MNEs (e.g. Meyer, Estrin, Bhaumik, & Peng, 2009; Meyer
& Nguyen, 2005). The second group focuses on the relationship between institutional
differences between home and host regions and the legitimacy of an operation under these
various institutional pressures (e.g. Bénassy-Quéré, Coupet, & Mayer, 2007; Kostova &
Zaheer, 1999). This group of literature argues that larger differences in institutions between
two regions may discourage FDI between them because potential conflicts between local
adaption and internal consistency create greater liability of foreignness (e.g. Bénassy-Quéré
et al., 2007).
However, both groups of studies have neglected the influence of home-region
institutions also provide valuable insights into firm's global strategy. Firm's foreign expand
strategy may be influenced by the home country in both direct and indirect ways. A direct
influence, for example, the image of the home country perceived by government and
consumers in the host country can be served as an advantage or liability for firm's global
strategy Cuervo-Cazurra (2011). Besides, the institutional distance between home and host
regions related to economic freedom, political influence or FDI restriction also affects the
destination of OFDI (Kang & Jiang, 2012). Chen et al. (2012) find empirical evidence in
Chinese OFDI. They show that as the development of more efficient institutions at home
reduces the institutional differences, Chines MNEs are more encouraged to invest in
developed markets.
As for the indirect influence of home country institutions on firm's global strategy are
mainly related to the firm capability that induced by home country conditions. For example,
the development of market-supporting institutions create the conditions for EMNEs to build
skills and capabilities in technology, branding, and management, which in turn enables the
firm to invest in developed markets (Chen et al., 2012). Moreover, Stoian (2013) finds that
the overall progress in home country’s institutional reforms, such as privatization, trade and
competitiveness of the home country and enhance the ownership advantages of EMNEs, thus
enabling firms to invest abroad.
Despite the growing interest in studying the role of institutional differences between
home and host countries in the context of OFDI (Peng et al., 2008), most existing research
has viewed institutions as one of the determinants of OFDI decisions rather than factors that
influence the extent to which MNE parent firms benefit from OFDI. Furthermore, most
studies have not examined institutions at the sub-national levels but consider within-country
institutions as a homogeneous entity, with rare exceptions (Chan, Markino, & Isobe, 2010;
Meyer & Nguyen, 2005). However, large-scale emerging economies with multiple
administrative regions, such as China and India, are characterized by diverse institutions
across sub-national regions (Liu, Lu, & Chizema, 2014). As a result, regional institutional
environments create conditions that form firm-specific resources and capabilities (Meyer &
Nguyen, 2005). Hence, it is important to consider institutional environment at sub-national
levels when investigating its impact on MNEs’ OFDI strategies in emerging markets (Chen et
al., 2012). This study aims to fill these gaps by explaining the role of home-region
institutions in EMNE parent firms as OFDI reverse spillovers receivers.
In innovation literature, comparisons of innovation performance across countries focus
on the structures and dynamics of national innovation systems (Lundvall, 1992; Nelson,
1993). However, as noted in Liu and White (2001), an aggregate level of analysis at the national level is associated with plenty of questions and criticism owning to the regional
disparities within a nation, especially in transitional and emerging economies. There is a
considerable literature proposing that innovation performance varies not just between
countries but also between sub-national regions (e.g. Acs, Anselin, & Varga, 2002; Fritsch,
2002). This is because knowledge generation and new technology production have a
tendency to cluster spatially (Li, 2009). Given the uncertainty, complexity and tacit form of
new knowledge, it cannot be fully articulated and may only be transferred through
face-to-face interaction and trust-based relations (Fu, 2008; Morgan & Cooke, 1998). Thus,
the knowledge and technical capabilities also tend to be geographically-bounded, and that
knowledge spillovers tend to be localized (Cooke, Gomez Uranga, & Etxebarria, 1997;
Cooke, Heidenreich, & Braczyk, 1998). That is to say, the sensitivity of the transfer of new
knowledge to distance provides an essential reason for the development of regional
innovation clusters (Cantwell & Iammarino, 2000). Therefore, spatial proximity could be
(Howells, 1999). This has inspired researchers to extend the innovation system framework to
the regional dimension, which gives rise to the concept of regional innovation system (RIS).
RIS theory is based on the assumption that location and spatial proximity matter for
innovation activities (Cooke et al., 1997). Cooke et al. (1998: 1564) define regional
innovation system as "systems in which firms and other organizations are systematically
engaged in interactive learning through an institutional milieu characterized by
embeddedness." Iammarino (2005: 499) adds that RIS is composed of "the localized network
of various actors and institutions in different sectors whose activities and interactions
generate, absorb, and diffuse new technologies within and outside the region." As a network
provides opportunities for exploiting new knowledge combination and creation, it triggers a
positive innovation cycle(Nightingale, 1998). The main groups of actors in a region that may influence the innovation activities are agents (e.g. inventors and entrepreneurs), public
research institutions, supportive services, and the regional workforce. It is the interaction, the
density, and the quality of the network between these actors that determine how innovation is
generated in a certain area (Fritsch, 2002). The recognition of the centrality of relationships
and the mutual influences among the actors in the system is indeed the key insight of the
constitutes an adequate approach to analyze the determinants of innovation performance in
the context of countries that cover huge geographical areas and feature substantial regional
disparities regarding economic and innovative capabilities (Fu, 2008; Yang & Lin, 2012).
Extant literature has investigated several determinants of regional innovation
performance. For example, Jaffe, Trajtenberg, and Henderson (1993) show that R&D
expenditures contribute the most to regional innovation output, which is measured by the
number of patents or new products launched at the regional level. Besides the increase in
R&D expenditures, Sleuwaegen and Boiardi (2014) find that an increase in high-quality
R&D staff that is regional intelligence is also a strong driver of patent growth. What's more,
noted by Cornett (2009), the organizational and functional aspects of knowledge-based
regional development policies are important factors that stimulate innovative activities since
they can foster knowledge dissemination, innovation, and entrepreneurship in local industrial
sectors.
In the past two decades, innovation performance in China has increased dramatically.
According to World Intellectual Property Organization (WIPO), the number of patent
applications per capital in China grew almost 13 times between 1995 and 2007 (Li, 2012).
third-ranked nation in patenting in the world. Hu and Jefferson (2009) find that in addition to
R&D intensification, foreign direct investment, a more patent-friendly legal environment, and ownership reform with more entry of non-state enterprises all significantly contribute to the patenting boom in China during 1995-2001.
However, there has been an increasing disparity in innovation performance across
Chinese regions (Li, 2009). Chung (2002) argues that the high degree of “coherence” and
“inward orientation” at the provincial level justifies that each Chinese region constitutes an
innovation system. Despite the increasing emphasis on RIS in China, several scholars have
stressed the need for more research into exploring the underlying determinants of innovation
disparity at the regional level (Furman, Porter, & Stern, 2002). Some studies have started to
address this call. For example, Li (2011) points out that an institutional change at the
provincial level, the implementation of patent subsidies programs, is one important facilitator
for regional patenting growth. Li (2009) also finds that inefficiency in the innovation process,
which determined by government support and the constitution of the R&D performers (firms,
universities, and research institutions), has a significant impact on the regional disparity in
innovation performance. Nevertheless, few studies have considered the effects of Chinese
how OFDI interacts with region-specific institutions. Hence, this study complements RIS
literature by illuminating how OFDI concerning country diversity and region-specific
institutions affect domestic, regional innovation performance in emerging economies.
3. Conceptual framework and hypotheses
In this section, we develop our hypotheses on how the diversity of OFDI country
destinations affects domestic, regional innovation performance, and how and what
region-specific institutions influence this relationship. Fig.1 shows our conceptual model.
Fig.1. Conceptual model.
3.1. OFDI country destination diversity and domestic, regional innovation
EMNE’s major motivation of OFDI is to access strategic assets such as superior
technologies and managerial practices from their counterparts from developed markets and
factors that can influence this reverse spillover effects is the extent to which EMNEs have the
opportunity to learn from foreign firms, while OFDI potentially brings up this opportunity
(Zhang et al., 2010). We further argue that the diversity of OFDI country destinations can
additionally contribute to the reverse knowledge spillover effects.
With greater diversity of OFDI country destinations, EMNEs are exposed to new and
diverse knowledge from multiple markets and cultural perspectives (Hitt, Hoskisson, & Kim,
1997; Li, Wu, & Zhang, 2012). As Zahra and George (2002) argue, exposure to an
environment with diverse technologies and management practices can facilitate firms’
innovation and openness. Moreover, knowledge diversity in the environment “provides a
more robust basis for learning because it increases the prospect that incoming information
will relate to what is already known” (Cohen & Levinthal, 1990: 131). Wijk, Bosch, and
Volberda (2001) show empirical evidence that the breadth of knowledge exposure has a
positive impact on a firm’s propensity to explore new and related knowledge. Therefore,
international diversification can facilitate innovation because it is considered as a channel for
EMNEs to approach a greater variety of innovative ideas, technologies and complementary
assets brought by their foreign counterparts through OFDI in global markets (Wu, Chen, &
Besides, knowledge diversity can also provide more opportunities for EMNEs to
recombine these technologies and practices to create new competitive advantages. Noted by
the innovation and knowledge management literature, recombining existing elements of
knowledge into new syntheses often brings about new knowledge creation (Katila & Ahuja,
2002; Kogut & Zander, 1992). As foreign counterparts in different countries bring various
technologies and management practices, the knowledge pool enhances regarding scale and
scope. This improves the possibility for EMNE parent firms to find new useful recombination
of these elements, hence facilitating new knowledge creation (Zhang & Li, 2010). From this
point of view, the greater the diversity of technologies and management practices brought by
foreign counterparts through OFDI, the greater the knowledge combination potentials, and
thus greater innovation potentials for EMNE parent firms.
More specifically, drawing insights from the literature on OFDI reverse knowledge
spillovers, we argue that a greater diversity of OFDI country destinations can enhance the
reverse spillover effects through the four channels discussed earlier.
First, given that EMNEs can observe and imitate a wider variety of technological
know-how and management practices brought by foreign counterparts from a greater
through the merger and acquisition of foreign firms, EMNEs can have face-to-face
communication with various foreign firms which is essential for transferring sophisticated
and tacit technologies or know-hows (Wu et al., 2016).
Second, as EMNEs build extensive business linkages with host upstream or downstream
firms in diverse countries, a greater variety of knowledge and technological know-how can
be transmitted back to EMNE parent firms (Li et al., 2016), which can further facilitate new
knowledge creation. For instance, EMNEs can directly interact with global competitors and
suppliers through the establishment of R&D or manufacturing bases in diverse foreign
countries, and thus timely obtaining technologies and market information which is helpful for
innovation (Wu et al., 2016).
Third, when employees of EMNE foreign subsidiaries from a greater variety of country
origins, they bring a wider variety of technological know-how and management practices to
the subsidiaries, in turn, EMNE parent firms (Li et al., 2016). EMNE parent firms in order to
benefit from the knowledge, they may conduct personnel exchanges by transferring
subsidiaries researchers to parent firms or sending researchers from parent firms to the
Fourth, EMNEs investing in diverse countries are exposed to more competitive
international markets than firms operating only in a domestic market. The competitiveness
will force these firms to innovate rapidly and update their technologies and techniques to
remain viable in competitive international markets (Hitt et al., 1997; Li et al., 2012). While
firms operating in domestic market may be isolated from such competition due to trade and
geographical barriers, diversified international firms are compelled to adopt best-practice
technologies to survive (Blalock & Gertler, 2004). From this perspective, a greater diversity
of OFDI country destinations enhances the competitive pressure faced by EMNEs, in order to
stay viable in highly competitive global markets, these internationally diversified firms will
invest more in innovation to achieve competitive advantages (Wu et al., 2016). Based on the
arguments above, our first hypothesis is thus:
Hypothesis 1. All else being equal, the diversity of OFDI country destinations is positively
related to domestic, regional innovation performance.
3.2. Moderating effect of intellectual property right protection
In previous hypothesis we argue that the diversity of OFDI country destinations can
intellectual property right (IPR) protection can be beneficial for EMNEs to transfer and
utilize the diverse knowledge acquired through OFDI.
It is often assumed that strong IPR protection limits the risk of expropriation and
opportunities of technological learning by local firms (Maskus, 2000). That is to say,
increased protection of IPR makes imitation for other domestic firms more
difficult (Branstetter, Fisman, & Foley, 2006). According to Chen and Puttitanun (2005), less
imitation means more incentive for the domestic innovating firm to invest more in innovation
because the technological advantages can be well protected. As we argued above, EMNEs
with a greater diversity of technical know-how acquired through OFDI have greater
innovation potentials. These EMNE parent firms are those that have the ability to develop a
patentable new technology. In a stronger IPR environment, technology appropriation can be
curbed by well-established contract enforcement and intellectual rights protection
mechanisms (Zhou & Poppo, 2010). Hence, local innovators are more encouraged to
innovate because in these regions the cost of knowledge expropriation is significantly
reduced and minimized (Shi, Sun, & Peng, 2012). Chen and Puttitanun (2005) also show
empirical evidence that strong IPR enforcement improves local firms' innovation in
In developed economies, firms can readily enforce IPR law because strong institutions
shield against illegitimate imitation (Coriat & Weinstein, 2002). While in emerging
economies, there is often a weak legal infrastructure and poor enforcement of property rights
and contracts (Siebeck, Evenson, Lesser, & Braga, 1990). What's more, unlike the typical
view that IPR protection within a given nation is similar across its regions, emerging
countries show wide cross-regional variations in the effectiveness of IPR enforcement. In
regions with strong IPR protection and implementation, firms are encouraged to invest more
in R&D and advanced technologies. By contrast, in regions with weak IPR enforcement,
firms face the risk of value expropriation and cannot use patents to appropriate returns from
innovation (Yi et al., 2015). In countries, like China, which have a short history of enforcing
IPR laws to protect innovation and that feature different regional institution and market
development, these cross-regional variations are particularly significant. In China, as a result
of administrative decentralization, provincial governments and regional authorities have
substantial judicial independence, and they frequently affect courts' judgments (Peck &
Zhang, 2013). Overall, we expect that EMNE parent firms in a strong IPR protection
OFDI, thus increasing their domestic, regional innovation performance. We formulate the
following hypotheses:
Hypothesis 2. The positive relationship between the diversity of OFDI country destinations
and domestic, regional innovation performance is positively moderated by IPR protection.
3.3. Moderating effect of market development
Marketization is defined as the degree of market-oriented mechanism development and
institutions in order to achieve more efficient market functioning (Fan, Wang, & Zhu, 2007;
Henisz, Zelner, & Guillen, 2005). Given that sub-national regions in large emerging countries
differ in the degree of marketization (Khanna & Palepu, 1997), we argue that the level of
market development in a region moderates the effects of the diversity of OFDI country
destinations on domestic, regional innovation performance.
First, a high degree of marketization implies a high level of market monitoring
mechanisms and legal systems (Shi et al., 2012). In regions with strong market development,
there are well-developed markets and technology transfer institutions so that institutional
uncertainty and transaction costs are reduced. Hence, EMNEs are more willing to transfer
various technologies and entrepreneurial experimentation back home (Yi et al., 2015). In
technology transfer, poorly developed markets and institutions pose challenges for EMNEs
and reduce their willingness to transfer diverse technologies back home (Tihanyi & Roth,
2002). Therefore, as well-established markets and institutions facilitate technology transfer
by EMNEs, there is greater availability of various knowledge resources that creates greater
potentials for innovation (Kafouros & Forsans, 2012).
Second, Zhang et al. (2010) propose that the extent to which firms can benefit from
diverse knowledge brought from foreign firms is contingent upon their capacity to learn from
foreign firms. That is a firm's absorptive capacity, "the ability to recognize the value of new
information, assimilate it, and apply it to commercial ends" (Cohen & Levinthal, 1990: 128).
Here we argue the level of market development in a region influences firms' ability to learn
from foreign firms (Meyer & Nguyen, 2005), and thus facilitates the absorptive capacity of
EMNEs.
Regions within emerging market with a high degree of marketization usually indicate
that the commodity and factor markets have achieved significant improvement (Peng, 2003).
Moreover, that certain market intermediaries and legal frameworks are similar to those in
developed economies (Peng, Sun, Pinkham, & Chen, 2009). This similarity means that
developed economies to achieve a competitive edge. Relying on absorptive capacity literature,
a firm's ability to use new knowledge elements depends largely on the firm's existing
knowledge stock (Cohen & Levinthal, 1990). Hence a certain level of knowledge overlap is
necessary for a focal firm to employ the knowledge stock of another firm (Tallman, Jenkins,
Henry, & Pinch, 2004). Therefore, EMNE parent firms from regions with high market
development are more capable of recognizing the value and contents of different knowledge
brought by foreign firms from various country origins.
Furthermore, the well-established institutions in a region also facilitate the exchange of
factors for value creation, which in turn helps firms to develop their absorptive capacity
(McEvily & Zaheer 1999). For example, a well-functioning labor market facilitates labor
mobility and accumulation of human capital so that domestic firms can recruit employees
from foreign firms, which in turn helps domestic firms acquire new and extend existing
innovative capabilities, thus facilitating their absorptive capacity. Based on the arguments
above, we hypothesize:
Hypothesis 3. The positive relationship between the diversity of OFDI country destinations
and domestic innovation performance is positively moderated by the degree of market
3.4. Moderating effect of international openness
International openness refers to an institutional context with a bunch of policy incentives
facilitating MNEs to enter or go outward in a focal country (Ortega & Peri, 2014). However,
although policies are set at a country level, they are implemented at a regional level (Meyer
& Nguyen, 2005), resulting in cross-regional variations in international openness within a
given country. In China, for example, although the government has actively supported OFDI,
the implementation of these policies varies across provinces. Some provinces opened earlier
than others for foreign investment. This makes international openness in China spatially and
structurally uneven (He, Wei, & Xie, 2008).
Again, drawing from the perspective of Zhang et al. (2010) that the extent to which
firms can benefit from diverse knowledge brought from foreign firms depends on their
absorptive capacity. We propose that variations in international openness across regions
moderate the effects of the diversity of OFDI country destinations on domestic, regional
innovation because the degree of international openness influences EMNE parent firms'
ability to utilize knowledge acquired through foreign markets.
Building on knowledge spillovers literature, many scholars have suggested that the
2013). These effects may be realized through various channels. First, through exposure to an
open trade regime, local firms can observe or learn through the imitation of foreign firms'
products and technologies and adapted by their R&D efforts (Blomström & Kokko, 1998).
Second, local firms may benefit from employee turnover whereby skilled employees from
foreign affiliates take a job in local firms carrying with them valuable knowledge. Third,
knowledge spillovers may be transmitted through upstream or downstream value-chain
linkages from foreign firms to local suppliers/distributors (Spencer, 2008). Fourth, great
international openness raises the level of competition in a region, forcing local firms to
become more efficient. To meet this competitive challenge, local firms in these regions are
likely to develop flexible strategies, learning practices and advanced technologies
(Blomström & Kokko, 1998).
Overall, international openness assists the integration of emerging countries into the
global economy (Lou & Tung, 2007). Indigenous firms in such environment are forced to
observe, learn from, and imitate the superior competencies of their foreign competitors
(Durán & Ubeda, 2005). As the existing stock and quality of knowledge of EMNE parent
firms will influence the extent to which knowledge reverse transfers from overseas markets
also be able to adapt and exploit those assets subsequently to be beneficial to their home
economies (Li et al., 2016). Thus these four effects will also strengthen EMNE parent firms
ability to learn and transfer knowledge from foreign markets, which enable them to make the
most of the resources they have accessed through their outward direct investment (Li et al.,
2016). Therefore, EMNE parent firms from a higher degree of international openness regions
are more capable of taking advantage of various technologies and management practices that
acquired through their OFDI from different country origins. We thus hypothesize:
Hypothesis 4. The positive relationship between the diversity of OFDI country destinations
and domestic, regional innovation performance is positively moderated by the degree of
international openness.
4. Methodology
4.1. Sample and data collection
We choose China as the empirical setting for this study of the effect of OFDI country
destination diversity on domestic, regional innovation for three main reasons. First, China is
an example of a large emerging economy and was the third largest source of FDI outflows in
the world by the end of 2015 (UNCTAD, 2016). Second, China is a country with significant
regions (provinces & municipalities) (Peck & Zhang, 2013), this makes it more appropriate to
conduct a regional level analysis instead of an aggregate national level. Third, the Chinese
Government has actively supported OFDI through its Go Global policy (Luo, Xue, & Han,
2010), suggesting that China can receive benefits from it.
In line with a large number of previous literature that study Chinese regional innovation
system, we choose the administrative provincial-level regions as the unit of analysis (Fu,
2008; Li, 2009; Yang & Lin, 2012). We use a balanced panel dataset for 30 provinces and
municipalities over the period 2011–2014. Tibet is excluded from the analysis because of the
limited availability of data.
We obtain OFDI data of each 30 Chinese regions from the Statistical Bulletin of China’s
Outward Foreign Direct Investment, compiled by the National Bureau of Statistics (NBS),
the Ministry of Commerce, and the State Administration of Foreign Exchange of China (Li et
al., 2016). The data for innovation and R&D are drawn from The China Statistical Yearbook
on Science and Technology and The Database of China Main Since & Technology Index
collected by the Ministry of Science and Technology of China (Li et al., 2016). The
region-specific institutional data are assembled from various issues of The China Statistical
and internally consistent because all firms in China are required to cooperate with the NBS
and submit their corporate information to the NBS (Chang & Xu, 2008). Hence, the statistics
published by NBS have been largely used by previous studies focus on Chinese strategy and
international business (e.g., Buckley, Clegg, & Wang, 2007; Chang & Xu, 2008).
4.2. Model specification
In line with prior studies of regional innovation performance, we base our model on the
knowledge production function proposed by Griliches (1979). This knowledge function
assumes that the outcome of successful R&D expenditure represents innovation and that
innovation is a function of the inputs to the R&D process (Cohen & Levinthal, 1989):
𝐼𝑁𝑁 = 𝑎×(𝑅𝐷𝐼𝑖)!
where
INNi = Innovation performance in region i
RDIi = R&D inputs in region i
Here we expand this basic model by including our explanatory and other control
𝐼𝑁𝑁𝑖, 𝑡 = 𝑎 + 𝛽1 ×𝑅𝐷𝐼𝑖, 𝑡 + 𝛽2 ×𝐷𝐼𝑉𝑖, 𝑡 + 𝛽3 ×𝐼𝑃𝑅𝑖, 𝑡 + 𝛽4 ×𝑀𝐴𝑅𝑖, 𝑡 + 𝛽5 ×𝐼𝑁𝑇𝑖, 𝑡
+ 𝛽6 ×𝐷𝐼𝑉𝑖, 𝑡 × 𝐼𝑃𝑅𝑖, 𝑡 + 𝛽6 ×𝐷𝐼𝑉𝑖, 𝑡 × 𝑀𝐴𝑅𝑖, 𝑡 + 𝛽6 ×𝐷𝐼𝑉𝑖, 𝑡 × 𝐼𝑁𝑇𝑖, 𝑡
+ 𝛽4 ×𝑍𝑖, 𝑡 + 𝜀𝑖, 𝑡
where
DIVi,t = OFDI country destination diversity in region i in period t
IPRi,t = Intellectual property right protection in region i in period t
MARi,t = Market development in region i in period t
INTi,t = International openness in region i in period t
Zi,t = Vector of control variables for region i in period t
εi,t is an independent and identically distributed error term
In studies of innovation performance, it is worth noting that there may exist reverse
causation that generates estimation problems (Zhang et al., 2010). This means that the
explanatory variables may affect innovation, but innovation may also influence some of the
explanatory variables. For example, OFDI may result in better domestic innovation
performance, but more innovative firms are also more likely to have OFDI (Li et al., 2016).
Therefore, the number of OFDI (relatedly, the diversity of OFDI country destinations) from a
to rule out this possible simultaneity and endogeneity issues, we lag the dependent variable
by one year to account for the fact that spillover effects take times to materialize and to avoid
possible endogeneity with the independent variables (Yi et al., 2015; Li et al., 2016).
4.3. Estimation methodology
We analyze the 4-year panel data by utilizing negative binomial models in generalized
estimating equations (GEE). GEE are an extension of generalized linear models applied to
longitudinal data (Liang & Zeger, 1986). We choose this estimation technique to estimate
more efficient and unbiased regression parameters relative to ordinary least squares
regression because of three reasons.
First, our dependent variable takes the form of count data. Consequently, ordinary least
squares regression is inappropriate since our dependent variable is not normally distributed
(McCullagh, 1984). Our dependent variable is in negative binomial distribution as a
goodness-of-fit test rejects the Poisson distribution assumption due to over-dispersion
(Gardner, Mulvey, & Shaw, 1995). Given that GEE approach facilitate regression analyses
on dependent variables that are not normally distributed (Nelder & Baker, 1972), we run
appropriate link function involves modeling the logarithm of the mean, this will be auto
specified by the model (Gardner et al., 1995).
Second, GEE accounts for autocorrelation by estimating parameters and standard errors
based on an estimation correlation derived from within-cluster residuals (Liang & Zeger,
1986). When data consist of repeated measures, there may be correlation within a subject
over repeated measures. Thus, researchers must account for the correlation within responses
when estimating regression parameters (Ballinger, 2004), in our case that would be the
unobserved factors influencing patterns in particular provinces over time. With the use of
GEE for time series data, the standard errors are likely to be autocorrelated over time (Greene,
1990).
Third, GEE is more robust than other panel data methods since it offers multiple
correlation matrix structures to best match the data (Liang & Zeger, 1986). As suggested by
Ballinger (2004), for data that are correlated within a subject over time with equal time
interval repeated measures, an autoregressive correlation structure (AR1) is an appropriate
correlation matrix because it sets the within-subject correlations as an exponential function of
this lag period. Hence, we specified our working correlation matrix as autoregressive
4.4. Variables and measures
4.4.1. Dependent variable
Domestic, regional innovation performance
Domestic, regional innovation performance is measured by the number of granted
patents in each province in each year (Fu, 2008; Paci & Usai, 1999). Patent counts are mostly
used measure in previous research since they identify innovation performance more
accurately than alternative measures such as "new product" sales (Wang & Lin, 2013). This is
because patents capture both product and process innovation (Fu, 2008), while "new
products" are often roughly defined, and in many countries like China they can be potentially
over-recorded by firms to receive subsidies (Li, 2009). Furthermore, the process of patent
registration guarantees the quality and public availability of data (Griliches, 1979). Overall,
patent counts provide a homogeneous and meaningful indicator of innovation performance
across countries (Malerba, Orsenigo, & Peretto, 1997).
4.4.2. Independent and moderating variables
OFDI country destination diversity
create the entropy of the diversity of OFDI country destinations at province level:
𝐷𝑖𝑣𝑒𝑟𝑠𝑖𝑡𝑦 = − 𝑆𝑖 ln(𝑆𝑖) !
!!!
where Si is the proportion of the number of countryi over the total number of OFDI countries
invested by each region in a year. N is the total number of different countries that invested by
each region in that year. Entropy equals zero if all OFDI from a province in a year have the
same country destination and it rises with the extent of the diversity of OFDI country
destinations in a province. This variable was calculated for each of the 30 provinces and
municipalities and was updated yearly.
We collect OFDI destination country data by each province over 2011 to 2014 from
“Cross-border Investment Firm List” provided by Department of Outward Investment and
Economic Cooperation, Ministry of Commerce, China (Marukawa, Ito, & Zhang, 2014). This
dataset contains 19,526 overseas affiliates and branches of Chinese firms (including
Foreign-invested or Joint-Venture firms) throughout 31 provinces and municipalities
approved by the Chinese government during 2011-2014. We categorize these data by the
regional origin of the investor and the approval date of OFDI. We then remove those
population of 17,720 firms that conduct OFDI over the period 2011-2014.
Intellectual property right protection
IPR efficiency is measured as the proportion of granted IP lawsuits over the total
application number of IP lawsuits in a region (Yi et al., 2015). We obtain IP lawsuits from
the website of the State Intellectual Property Office of China. A higher ratio of the IPR
enforcement implies a stronger regime of IPR protection in the region.
Market development
Following Yi et al. (2013) and Hong et al. (2014), we use a marketization measure
developed by the National Economic Research Institution (Fan et al., 2007). This is a
comprehensive composite index that provides an overall assessment of marketization in each
Chinese province (Shi et al., 2012). The overall index is accessed via five key areas:
government and market roles; development of non-state-owned enterprises; development of
commodity markets; development of factor markets; and development of free market
institutions and a legal environment. A higher score indicates a higher level of market
development.
International openness is an institutional arrangement with a set of policy incentives
facilitating MNE to enter or go outward in a focal country. In line with numerous studies (e.g.
Ortega & Peri, 2014; Yi et al., 2015), international openness is measured by the ratio of
imports and exports to GDP in a province. A higher ratio shows a greater degree of
international openness.
4.4.3. Control variables
We include five control variables. First, building on the knowledge production function
proposed by Griliches (1979), we include R&D inputs as one of the control variables. In line
with previous studies of innovation process at the sector (e.g. Li, 2011) and regional levels
(e.g. Fu, 2008), we use R&D intensity to measure the R&D inputs. The R&D intensity is
calculated by dividing the R&D expenditure by the GDP of the region in a particular year,
thus taking into account the relative sizes of the regional economies.
Second, to control the development potential and regional demand for innovation, we
include the regional GDP growth rate as one of the control variables (Ouyang & Fu, 2012).
This is because we expect regions with faster economic growth have stronger innovation
that firms will also have the desire and the expertise to produce those products (Li et al.,
2016).
Third, given that foreign technology has been a critical source of advanced knowledge
for many developing countries to increase their innovation capabilities (Yang & Lin, 2012), it
is important to control for this potential determinant of innovation performance. Moreover,
over the period covered by this study, China was a major buyer of foreign technology (Li,
2011). We thus expect those regions where foreign technology purchase account for high
proportions of technology transaction to exhibit higher levels of innovation (Li et al., 2016).
Fourth, drawing insights from previous studies that state-owned enterprises (SOEs) may
be less interested in innovation than private enterprises because they generally operate in
protected industries like energy and national defense (Lin, Cai, & Li, 1998). We thus employ
the proportion of SOEs capital investment over the total value of capital investments in each
region to control its negative impact on innovation performance (Li et al., 2016).
Last but not least, as argued in our conceptual framework, regional absorptive capacity
is important to facilitate the realization of the benefits of OFDI for domestic innovation. In a
similar vein, scholars suggest that regions with a certain scientific base and amount of R&D
reverse transferred from overseas markets (Deng, 2007; Durán & Ubeda, 2005). Thus,
spillovers to domestic economies are more likely to take place in such regions. Hence,
assuming that highly educated people are typically the ones to undertake the innovation
process, we measure the absorptive capacity of each region by calculating the ratio of R&D
staff to the total employees of a region in a given year (Li et al., 2016). Table 1 provides
detailed definitions of each of the variables. Table 1
Description of variables.
Variable name Acronym Operationalization
Innovation performance INNit Granted patents of region i in year t
Diversity of OFDI country DIVit Formula (3) of region i in year t
Intellectual property right IPRit Granted IP lawsuits over the total application number of IP lawsuits of region i in year t
Market development MARit Marketization index (Fan et al., 2007) of region i in year t
International openness INTit The ratio of imports and exports to GDP of region i in year t Research intensity RDIit R&D expenditure/GDP (percent) of region i in year t
GDP growth rate GDPit GDP growth rate (percent) of region i in year t
Foreign technology purchase FTECHit Foreign technology purchase value over the total value of technology transaction in region i in year t
State sector’s share of output SOEit The share of capital investment of local SOEs' over the total of region i in year t
5. Findings
The number of granted patents nationally in China has risen over the 2012-2015 period,
from 1,163,226 in 2012 to over 1,596,977 in 2015. However, this national increase has not
been reflected uniformly across the Chinese regions, better innovation performance regions
have been concentrated in several coastal provinces (e.g. Jiangsu, Zhejiang, Guangdong).
Similarly, there has been a marked increase in OFDI flows from China over the period,
rising radically from US$ 74.65 billion in 2011 to US$ 123.12 billion by 2014. The most
active regions are the coastal provinces and municipalities (e.g. Guangdong, Jiangsu,
Zhejiang, Beijing, and Shanghai). Moreover, coastal provinces have invested in more diverse
foreign countries over the period (e.g. Shandong, Jiangsu, Zhejiang).
Table 2 exhibits mean, standard deviation, and the correlation matrix for all the variables.
The positive and significant correlation between dependent and independent variable
supports our hypothesis that regions with higher innovation performance have higher OFDI
countries diversity. Most of the correlations between the explanatory variables are small,
indicating that multicollinearity is not a serious concern in the estimation. Although the
correlations between market development and international openness (0.69), research