0
Joint innovation: Born globals from emerging
economies and their partners
By
Nikoleta Minova
Student ID: 11439890
Date of submission: January 26 2018
MSc Business Administration: International Management
Master Thesis
University of Amsterdam
First supervisor: Dr. Vittoria G. Scalera
1 This document is written by Student Nikoleta Minova who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are 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 completion of the work, not for the contents.
2 Content
1. Introduction ... 3
2. Literature review ... 7
2.1. The Born global ... 7
2.1.1. Born globals from emerging economies ... 9
2.2. Traditional MNEs ... 11
2.3. Alliances ... 12
2.3.1. Alliances in emerging economies ... 15
2.4. Innovation in alliances ... 17 3. Research question ... 18 4. Theoretical framework ... 19 5. Hypotheses ... 20 6. Method ... 25 6.1. Data sample ... 25 6.2. Variables ... 29 7. Statistical analysis ... 34 7.1. Regression analysis ... 37
8. Discussion and conclusion ... 41
8.1. Managerial and policy implications ... 44
8.2. Theoretical implications ... 45
8.3. Limitations and future development ... 46
9. Acknowledgements ... 48
List of references ... 49
List of figures Fig. 1: Theoretical framework ... 19
Fig. 2: Interaction term plot ... 41
List of tables Table 1: Industrial distribution of the MNEs in the dataset ... 27
Table 2: Geographical distribution of the companies in the dataset ... 28
Table 3: Variables definitions and coding ... 33
Table 4: Descriptive statistics and correlations ... 36
3 Abstract
Born globals from emerging economies increasingly capture the attention of scholars and practitioners alike. Their international success defies the fact that they start out from countries with low economic development and that they generally lack experience. This study addresses the shortage of research on born globals from emerging economies, focusing on their participation in international joint ventures (JVs) and the changes their partners incite in their innovation capabilities. The success of born globals depends on both developing innovative products (Cavusgil & Knight, 2015) and the alliances they enter to compensate for asset parsimony (Freeman, Edwards, & Schroder, 2006). The theoretical framework includes two partner characteristics as independent variables: JV form, horizontal or vertical, and traditional MNE (multinational enterprise) country of origin, developed or emerging. The analysis uses patent counts as the dependent variable to operationalize company-level innovation output. Moreover, the dataset consists of 220 JVs between born globals from emerging economies and traditional MNEs coming from developed or emerging economies formed between 2000 and 2015. Hypotheses 1 and 2 test if horizontal JVs and traditional MNEs from developed markets affect JV innovation output positively. Hypothesis 3 examines the moderating effect traditional MNE country of origin has on the relationship between JV form and JV innovation output. The statistical analysis supports all three hypotheses. Overall, the findings suggest that horizontal JVs and traditional MNEs from developed markets have a stronger positive influence on innovation for JVs with born globals from emerging economies in comparison to their alternatives, vertical JVs and traditional MNEs from emerging economies. Further, traditional MNEs from developed, rather than emerging countries, have a stronger positive effect on the relationship between JV form and JV innovation output.
Keywords: born global; emerging economy; developed economy; strategic alliance;
4 1. Introduction
In 2008 The Economist belatedly announced the arrival of ‘a new breed of multinational company’ (The Economist, 2008). The so-called “new boys”, according to the article, are multinationals from countries like India, China, and Brazil. Western economies are slowing down and emerging ones are already becoming major players. For reference, China has the highest GDP in PPP (Purchasing Power Parity), overtaking the USA, while India, Russia, and Brazil all rank in the top ten for 2016 (World Bank, 2017). Considering that in 2008 the four countries were also in the top ten for the same indicator (World Bank, 2017), the author’s surprise is unwarranted. In addition, among the multinational enterprises from these emerging economies, there is an even more interesting subset of companies that the world has overlooked – the born globals. Examples of such companies coming from China are the Alibaba Group, which is among the biggest online retailers worldwide, and the successful smartphone manufacturers Xiaomi and OnePlus. Consequently, I have to distinguish between two types of multinational enterprises depending on their pattern of internationalization. Throughout this thesis, I will refer to companies that approach internationalization gradually, step by step, and by accumulating experience and knowledge about the foreign markets (Johanson & Vahlne, 1977) as traditional MNEs and companies that engage in internationalization from their onset or shortly after as born globals (Knight & Cavusgil, 2004). This “unique breed” (p. 124) of firms defy the traditional patterns of internationalization with the scope and speed of its cross-border expansion (Knight & Cavusgil, 2004).
A considerable part of the IB literature is dedicated to multinationals from the developed world, whereas research focused on emerging economies is incomplete. Going further on the matter of born globals, some scholars maintain that they originate explicitly from developed economies (Guillen & Garcia-Canal, 2009). While this is at best highly debatable, most research on the topic does concern predominantly developed economies. Consequently,
5 the following literature overview concluded that academic research on born globals is a still new and insufficiently developed area, especially their knowledge-based capabilities in the context of inter-firm partnerships. The literature is mainly concerned with firms’ motives for entering an alliance quoting reason like entering foreign markets and overcoming resource constraints, but it does not follow through after the formation of the alliance. It is the same with partner selection criteria, upfront managers identify criteria they consider when choosing a partner (e.g. “technical capabilities”, “special skills you can learn from partner”, “willingness to share expertise”, etc.) (Dacin, Hitt, & Levitas, 1997, p. 10) based on assumptions and expectations, not on empirical analysis.
In sum, little is still known about emerging markets born globals and their innovation and knowledge capabilities, especially in the context of alliances. It is true that due to their size, small and medium-sized, born globals are well suited for operating abroad in terms of their flexibility, adaptability and exceptional learning capacity. However, before turning their international expansion into fuel for innovation, they have to secure the tangible resources necessary for competing on the global market – sufficient level of productivity to face the increased demand and financial resources to shoulder the additional exporting costs (Salomon & Shaver, 2005). This is in line with one of the tenets of the catch-up literature, which states that companies from emerging economies use international expansion as a “springboard” to acquire resources they lack instead of taking the time to amass them through path dependence growth (Luo & Tung, 2007). The framework I propose contributes to the literature of firm-innovation and alliances in the context of born globals by adding insights about the way born globals and traditional MNEs interact. This research aims to answer the question of how partner characteristics affect the alliance innovation output in born globals from emerging economies. For this study, I propose the following independent variables that affect innovation within an alliance: horizontal or vertical (Ellis & Mayer, 2009) and partner firm country of origin,
6 developed or emerging. Moving forward, the term innovation output denotes the products or services resulting from the cooperation within the alliance and their degree of novelty. As a method to operationalize firm innovation, scholars suggest patents. Multiple academic works use patents as proxies for innovation either directly or by transforming them into different patent-based measures (Lanjouw & Schankerman, 2004; McAleer & Slottje, 2005; Mowery, Oxley, & Silverman, 1996). Further, they are well documented and reliable since patent law regulates their issue.
While the theoretical scope of this study is equity strategic alliances, empirically it covers JVs. The analysis uses cross-sectional data for 220 international JVs between born globals from emerging economies and traditional MNEs from both developed and emerging countries. Each observation includes firm-level details for both parties of the alliance that help categorize them as traditional MNEs or born globals, as well as information on country of origin and industry affiliation related to the two independent variables. Furthermore, the dataset includes JVs founded between 2000 and 2015, ranging across 15 different industries. In order to test the hypotheses, 4 Negative Binomial Regression models are run, starting with the control variables of the research and progressively adding the two independent variables and one interaction term.
This study can potentially contribute to the current academic literature by addressing an underdeveloped but trending topic, namely born global firms. The theoretical framework and empirical approach I develop here aim to bridge the current understanding of born globals with what we know about alliances in order to reach new insights. Porter links innovation and constant improvement to global competitiveness, which is why every variable that affects it is worth examining (Porter, 1990). Assessing firm innovativeness through patents can help identify the direction of knowledge-transfer between the partnering firms. Should management be worried about loss of proprietary knowledge or focus on more proactive learning? In
7 addition, both born globals’ managers and those that collaborate with them will be able to make informed decisions in terms of partner choice and structure of the alliance before committing to a partnership as well as assessing its success afterward. On a national level, governments can enact policies encouraging domestic companies to partner up according to country of origin, industry or position along the value chain.
Finally, this thesis follows a clear and straightforward structure. The following section is the literature overview that includes three subsections each providing theoretical background for a key element: born globals from emerging economies, traditional MNEs, innovation and alliances. Parts 3 and 4 present the research question and theoretical framework. Section Method explains the choice of data sample, moderator, and control variables, and is followed by the statistical analysis. The thesis ends with the Discussion and conclusion section covering the current study’s expected theoretical and practical contributions.
2. Literature review
The following sections will provide an overview of the relevant academic literature as well as working definitions of the key elements of the framework – born globals, traditional MNEs, innovation, and alliances.
2.1. The Born global
The brief overview of the extant literature on international business reveals a lasting interest in the phenomenon of born globals. Although Ghemawat (2003) describes the world as semi-globalized, rather than truly global, there is an evident trend towards integrated markets and deregulation of international business (Chetty and Campbell-Hunt 2004). The born global cannot exist in a world of self-sufficient national markets without intensive international trade. To thrive, these companies need a business environment characterized by global connectedness along with progressively homogenous worldwide demand, technological developments,
8 efficient and cost-effective logistics, formation of a global middle class (Cavusgil & Knight, 2015). The widening cross-border networks of suppliers and customers and the current trend of cross-border market integration fosters a strong expectation of an increase in numbers and importance of the global start-ups, International New Ventures, and born global firms as a whole.
In this research I use Knight and Cavusgil’s broad definition of born globals as “organizations that, from or near their founding, seek superior international business performance from the application of knowledge-based resources to the sale of outputs in multiple countries” (2004, p. 124). Across the IB literature, there are conflicting views as to within what period after inception must internationalization take place, for a firm to be considered born global. From the explicitly short term of two years (McKinsey & Company, 1993) or three years (Knight and Cavusgil 2004), to the mid-term view of eight years (McDougall, Shane, and Oviatt 1994). This research focuses on the particular knowledge-based and innovation-creation capabilities as the key characteristics of born globals, not as much on pinpointing when exactly their cross-border expansion takes place.
Chetty and Campbell-Hunt (2004) summarize that born globals are driven to enter multiple markets quickly in order to capitalize on the potential of an innovative product. Due to their small size, they focus on a narrow product-market scope, and extensively use networks of business partners (Chetty & Campbell-Hunt, 2004). It is important to note here one of born global firms’ defining traits, namely, the fact that they manage to internationalize under conditions of asset parsimony (Cavusgil & Knight, 2015). Asset parsimony reflects born globals’ lack of traditional strengths which take time to amass including company resources and experience, believed to be the stepping stone for international expansion (Johanson & Vahlne, 1977). By contrast, born globals fall back on strong entrepreneurial orientation, persistence, innovation and differentiation (Cavusgil & Knight, 2015).
9 According to Knight and Cavusgil (2004), born globals owe their success to several key specific organizational capabilities concerning the organizational culture and business strategies born globals employ. In terms of organizational culture, what makes born globals different from traditional MNEs are their unique entrepreneurial outlook and proactivity in pursuing foreign markets. In addition, strong marketing orientation explains born globals’ focus on superior customer value creation, which helps them stay competitive in the global marketplace in the face of increasingly better-informed and more demanding customers. The authors also identify the four most important business strategies, global technological competencies, unique products development, quality focus, and leveraging foreign distributor competencies, that enable born globals to internationalize early and compete efficiently in the global business.
Among the born globals, there is another special group of companies that defy the classic IB tenets on international expansion. These are the early internationalizers coming from emerging economies as opposed to developed ones. The following section looks into the specific characteristics that discern the two types of born globals and examines the specific challenges born globals from emerging markets face.
2.1.1. Born globals from emerging economies
Another important characteristic of the born global is its home country. Firms from developed economies differ considerably from their emerging economy counterparts from the type of ownership advantages they possess (Dunning, 2000) to the way they internationalize (Johanson & Vahlne, 1977; Luo & Tung, 2007). Emerging economies are characterized by low levels of income, but rapid pace of development and market liberalization government policies (UNCTAD, 2006; Wright, Filatotchev, Hoskisson, & Peng, 2005). In addition, they are associated with institutional voids (e.g. weak intellectual property rights) underdeveloped
10 capital markets and, in some cases, political risk (Luo and Tung 2007). Firms coming from these countries usually lack tangible resources – finances, plant, equipment, and human resources (Knight and Cavusgil 2004). Guillén and García-Canal (2009) assume that born globals originate from developed countries because of the intangible assets that allow them to internationalize. However, this applies also to born global firms from emerging economies. Further, emerging-economy born globals are interesting not only because they are understudied. These companies compete on the international business scene with their own intangible assets, but need to overcome the dual inherent disadvantage of being young and coming from an underdeveloped country.
Extant literature provides several concepts, which combined provide a good explanation of how emerging economy born globals overcome the dual disadvantage of young age and country of origin. First, in his 2012 article Ramamurti explains the controversial existence of multinationals from emerging economies with special ownership advantages such as having a good understanding of local customer needs, the ability to do business in unfavorable market conditions, to make products at ultra-low cost, and to develop products that are “good enough” with the right price/feature ratio (Ramamurti, 2012). Second, as mentioned in the previous section, born globals employ higher order capabilities to gain competitiveness in the global marketplace. Making use of knowledge-based resources, innovativeness, superior entrepreneurial orientation and proactivity when it comes to internationalization (Cavusgil & Knight, 2015; Knight & Cavusgil, 2004) is how born globals manage to survive against their larger and more experienced competitors. Finally, born global companies do not shy away from approaching international partners in creating alliances and networks of partners (Freeman et al., 2006) as a way to mitigate their inherent weaknesses.
Born global companies and SMEs looking to internationalize share a number of similarities but also constraints. In their work on SME survival, Lee, Kelley, Lee and Lee
11 (2012) maintain that expanding into foreign markets and entering into RnD alliances increases their probability of survival. However, both actions bring additional levels of complexity for the SME that inherently lacks managerial experience. In terms of internationalization, SMEs, and by extension born globals, are at a disadvantage for several reasons including the SME-specific liability of smallness and the more general liabilities of newness and foreignness. However, by seeking out new markets, SMEs can escape severe home competition by capitalizing on overlooked opportunities abroad like market niches or markets that larger companies have not entered yet (Lee et al., 2012), the additional information inflow can also spur product innovation (Golovko & Valentini, 2011).
This section provides an overview of the relevant literature on born globals from emerging economies, while the following section will focus on their exact opposites in terms of globalization pattern – large MNEs originating from developed countries. Dubbed in this thesis traditional MNEs, it is important to look into these entities as they add meaningful context to the concept of born globals by providing a basis for comparison. In addition, the current research studies the relationship between born globals from emerging economies and traditional MNEs, so the scope and key characteristics of both terms need to be defined.
2.2. Traditional MNEs
The subjects of this study are two types of companies: born globals coming from emerging economies and “traditional” multinationals coming from developed countries. This section explains what I will refer to as traditional MNEs and why it is important to differentiate between them and born globals from emerging economies.
First, the term traditional MNE starts with the multinational enterprise, “a firm that owns and controls activities in two or more different countries” (Buckley and Casson 2009, p. 1564). Second, it is important to state that traditional MNEs come from developed countries as
12 this has major implications for their firm-specific and country-specific advantages. For example, developed countries have well-functioning formal institutions, modern infrastructure, stable economy and efficient goods, labor and financial markets (Schwab, 2017). Firms originating from developed economies also enjoy country-specific advantages such as country of origin effect, adding to their exceptional marketing capabilities in creating strong global brand names (Sheth, 2011). In addition to these advantages, most traditional MNEs also share disadvantages stemming from their size and age. Large established companies on the global market tend to be less flexible and slow in their decision making which makes it hard to keep up with market trends and changing consumer preferences (Knight & Cavusgil, 2004). In addition, red tape and taller hierarchies add to the difficulties in developing innovative products and marketing them on time.
The greatest difference between traditional MNEs and born globals, however, probably comes from the way they internationalize. As opposed to internationalizing from the get-go, traditional MNEs adopt the gradual internationalization process of the Uppsala model (Johanson & Vahlne, 1977). In the Uppsala model, cross-border expansion happens in stages. By first utilizing a low-commitment non-equity entry mode like indirect export, the firm goes through experiential learning in order to gather information and know-how about operating in the foreign market (Johanson & Vahlne, 1977). Although slow, this approach greatly reduces the negative effect of liability of foreignness and by increasing the firm’s level of involvement incrementally, makes it less costly to pull out of a market especially during the early stages.
2.3. Alliances
As mentioned in the previous sections, born globals from emerging economies have to overcome two types of obstacles to international expansion: stemming from their youth and stemming from their home country. Such companies usually lack tangible resources – finances,
13 plant, equipment, human resources (Knight & Cavusgil, 2004). In addition, emerging economies are associated with institutional voids (e.g. weak intellectual property rights) underdeveloped capital markets and, in some cases, political risk (Luo & Tung, 2007). One viable way to overcome both types of constraints is to engage in strategic alliances that range from purely contractual agreements to equity agreements like joint ventures.
Gulati (1998, p. 293) defines alliances rather broadly as “voluntary arrangements between firms involving exchange, sharing, or development of products, technologies, or services”. However, there are contrasting views in the literature about the scope of the term. On the one hand, Lin and Darnall (2015) include in the term alliances ownership, contractual and licensing agreements. They go on to explain that ownership agreements concern the commitment of equity, like in joint ventures and minority equity alliances, while contractual agreements warrant less interdependence - joint RnD, production or marketing and promotion activities. On the other hand, in their hierarchical model of entry modes Pan and Tse position alliances strictly under non-equity contractual agreements, at the same level as licensing and RnD (Pan & Tse, 2000). Moreover, joint ventures fall under equity modes of entry further separating them from the notion of an alliance. In addition to ownership structure, other authors distinguish between vertical alliance integration, with partners along the value chain such as suppliers or distributors, and horizontal, among competing firms, (Ellis & Mayer, 2009; Gulati, 1998).
Taking the resource-based view of the firm, Das and Bing-Sheng (2000) hold that alliances are successful when partners’ resources either supplement or complement each other. Going back to born globals’ sources of competitive advantages and disadvantages, I can theorize that the alliances they engage in are based on resource complementarity. More precisely, born globals do not seek to amass a critical volume of any resource, since they operate under asset parsimony (Cavusgil & Knight, 2015) and would not have much to
14 contribute in this sense to an alliance. On the contrary, born globals seek the support of their partners to access resources they lack in order to fully exploit and also enhance their firm-specific capabilities (Mowery et al., 1996). Such capabilities are entrepreneurial orientation, vision and innovativeness (Cavusgil & Knight, 2015). In this line of thought, a born global enters into an alliance to execute its firm strategy using resources it lacks, and to enhance intangible assets like managerial experience through inter-firm learning.
Freeman et al. (2006) further support the vital importance alliances have for born globals. Their article synthesizes the inherent constraints to early and rapid internationalization they face as poor access to economies of scale, lack of financial and knowledge resources and risk aversion (Freeman et al., 2006). The authors conclude with five strategies to overcome them and two of them center on the benefits of intra-firm collaboration. First, alliances mitigate born globals’ poor access to economies of scale by allowing them to tap early into many and/ or large markets, to outsource production to suppliers, and improve sales volumes by working with prominent distribution partners. Second, collaborative partnerships help firms access key markets at low cost thus sharing the cost of expansion. In addition, developing networks promotes access to valuable intangible resources like knowledge, including tacit, useful in new product development (NPD). Third, one of the most common reasons to form a partnership, not only within IB, is risk sharing, especially when the partnership is built on trust. Freeman et al.’s (2006) article shows that firms have a myriad of reasons to enter into an alliance and that it actually happens often. Because of the all-around effect alliances have on their participants, their impact on the foundations of a company’s core competitive advantage is especially important, for born globals these are innovation capabilities. Their impact depends also on the ownership structure of the alliance. Das and Bing-Sheng make the proposition that minority equity alliance will be more beneficial to a partner firm when its primary resources are knowledge-based and its alliance partner’s – property based (Das & Bing-Sheng, 2000).
15 However, in the same situation, the property-based resource partner will prefer an equity joint venture (EJV). Implementing this reasoning, born globals will prefer minority equity alliances when their partners will prefer EJVs. To sum up, for a large well-established company a bad choice of a partner or a type of partnership can result in financial or reputational damage. By contrast, for a born global firm such a mistake can result in bankruptcy.
2.3.1. Alliances in emerging economies
SMEs in emerging economies have several important reasons to enter into an alliance, be it domestic or international. Mesquita and Lazzarini (2008) argue that, ultimately, SMEs enter into such partnerships in order to secure “collective resources” and to overcome home country weak institutional and infrastructural environment, which enables them to compete on the global market. Hitt et al. add the need to improve legitimacy in the eyes of consumers and investors (Hitt, Ahlstrom, Dacin, Levitas, & Svobodina, 2004). Further, Mesquita and Lazzarini (2008) study two types of relationships between domestic partners in emerging economies, namely, horizontal and vertical. On the one hand, horizontal relationships include SMEs from the same industry or producing complementarity products, possibly competitors. On the other hand, vertical relationships are natural, not between rivals, but rather between companies that specialize in different activities along the value chain. Their paper states an important distinction between the two types of relationships. While vertical alliances result in higher manufacturing productivity, horizontal alliances boost collective sourcing of recourses and higher rates of innovation. Although these findings refer to domestic alliances, they focus the attention to an important distinction in the pattern of alliance. How would SMEs from emerging economies that internationalize early fare in this respect? It is worthwhile to study the effect of the type of relationship between alliance partners in the case of born globals and their partners further because of the different implications of an international setting and the impact it has on company innovation. It is interesting to note that in this particular relationship
16 innovation is both the foundation of SMEs competitive advantage and the reason why the partner firm would enter into an alliance.
Partner selection in an international setting is a daunting but crucial task because of country-specific differences such as culture, infrastructure, economic development and policies (Dacin et al., 1997). In their 1997 study, Dacin et al. look at firms from newly industrialized and traditional industrialized countries emphasizing on differences in their motivations to enter alliances and their expected benefits. In addition, the study connects level of economic development with motives to enter into an alliance. Firms from developed countries seek new markets to exploit while their developing country partners aim to gain access to advanced technology, opportunities for export and international alliance experience (Dacin et al., 1997).
Another research on alliance partner selection, specifically focused on emerging economies, connects a country’s institutional environment to short-term or long-term strategic orientation of managers and partner choice criteria (Hitt et al., 2004). Again, the authors point out the differences the home environment, emerging or developed economy, makes. For firms facing the challenges of operating in an unstable economic and market environment and aiming to remedy them by entering into an international alliance, the following selection criteria are outlined as key: technological and managerial capabilities, intangible assets, complementary capabilities, unique capabilities, and prior alliance experience. With the exception of complementary capabilities and prior alliance experience, the more stable the institutional environment, the greater the weight of these criteria. This is so because according to Hitt et al. (2004), institutional stability instigates long-term strategic orientation from managers, which influences their partner selection criteria.
The two studies on partner choice criteria commented above (Dacin et al., 1997; Hitt et al., 2004), however, do not address the impact country origin has on the alliance outcome.
17 Following the notion that the home country influences the collaboration between the alliance partners, this research will analyze the effect of country of origin on firm innovation.
2.4. Innovation in alliances
Knight and Cavusgil’s definition for born globals links firm performance to innovation as their sole provider of competitive advantage on the international markets. Born globals are essentially challengers of the status quo and of the industry incumbents. This view of born globals mirrors Schumpeter’s “creative destruction” perspective on innovation (1976), where innovation triggers a process of discarding the old for the new in terms of technology, processes, products. Coveted as one of the most important research issues in business today (Hauser, Tellis, & Griffin, 2006), innovation on the firm level can be defined as “the process of bringing new products and services to market” (Hauser et al., 2006, p. 687). It refers to the idea-generation capacity of a firm to develop novel products, to market them internationally and to creatively overcome difficulties (Knight & Cavusgil, 2004).
Academic works on innovation share a common problem, ultimately coming from the abstractness of the term. On the one hand, some authors propose to study innovation through different proxies or using secondary data. For example, knowledge-transfer and technological capabilities (Mowery et al. 1996); learning, replication, and creation (Wang and Nicholas 2017); experiential and inter-organizational learning (Bruneel, Yli-Renko, & Clarysse, 2010). On the other, because it is difficult to operationalize phenomena like knowledge, learning, and innovation, most researches in the sphere are qualitative in nature, often including small samples of cases.
Innovation is a central reason for companies to collaborate as evident from the increase in open innovation practices and RnD joint ventures (Kwak, 2004; Sampson, 2007). The topic is still vastly understudied but there are some academic works that shed light on the mechanics
18 of this type of inter-firm collaboration. For example, an empirical study establishes a noteworthy relationship between partners capability diversity (Sampson, 2007). Sampson maintains that alliances reach high levels of innovation when partners are moderately diverse. In other words, they still need a basis of common stock of knowledge to make collaboration possible, but diversity ensures that they have enough to learn from each other (Sampson, 2007). This conclusion is important for partner selection choice on company level but also for a pre-screening of the partner’s country and industry, which can also indicate similarity in capabilities and technological level.
The SME literature makes use of the term “learning by exporting” to denote the positive impact exporting to foreign markets has on firm productivity and innovativeness. However, by simply putting themselves in a situation where in order to succeed the firm has to step up is not enough. Salomon and Shaver (2005) rightfully note that except for having the capacity to learn, firms need a sufficient level of productivity as well as mechanisms to cope with the additional costs of exporting. This is a viable concern especially for SMEs, which are well suited for this way of learning and competing, but face considerable resource constraints.
3. Research question
The elements discussed in the previous paragraphs form an interesting focus of research. For newly founded and newly international firms like the born globals, committing to an alliance may seem like a sound decision. However, scholars as well as managers need empirical evidence to clearly understand which alliance partner characteristics impact born globals and how. This research addresses the relationship between the innovativeness of born globals from emerging economies and their international alliance partners, formulated into a research question as follows:
19
How do partner characteristics affect the alliance innovation output in born globals from emerging economies?
The study looks at international alliances and examines the relationship between partners along the value chain or within the same industry scope (horizontal vs. vertical alliances), and home country of the traditional MNE in the alliance, developed or emerging, and how these relationships impact the alliance innovation results. In addition, the research examines the moderating effect of partner country of origin on the relationship between horizontal/ vertical alliance and innovation output.
4. Theoretical framework
One way to examine a company’s innovativeness is by the products and services it markets, or the degree of novelty of its output. In turn, product or service novelty can be observed through a firm’s patent portfolio. For example, in their research on inter-firm knowledge transfer Mowery et al. (1996) calculate a measure they call cross-citation rate to establish the relative importance of a partner’s knowledge contribution to the alliance. This way cross-citation rate tracks the path dependence of a product’s intangible inputs. Another study examining the effects of exporting on innovation also uses patent application counts
Innovation Output Horizontal vs. Vertical Alliance Developed vs. Emerging Market Partners
20 along with product innovation counts to assess changes in firm innovation (Salomon & Shaver, 2005).
5. Hypotheses
SMEs from developing economies that wish to increase their international competitiveness by engaging in alliances can choose between vertical or horizontal ties, both of which create different dynamics between the partners (Mesquita & Lazzarini, 2008). First, vertical alliances are easier to incorporate into the production process, as one agent’s outputs become the other’s inputs, much like the internal working of a company. Activities organized in a value chain are connected sequentially, creating interdependent relationships (Mesquita & Lazzarini, 2008). In such partnerships, each party is solely responsible for a specific task or stage of the overall process (e.g. production of components, assembly) but is dependent on the performance of its preceding stages (Mesquita & Lazzarini, 2008). Since the partner companies possess distinctive capabilities and their contributions are based on complementarity, in most cases they do not directly compete on the market. Consequently, if the possibility of conflict of interest is low, partners will be more willing to share proprietary knowledge and expertise since they will not be as guarded against possible opportunistic behavior. This is also in line with Sampson’s (Sampson, 2007) findings that diversity in partners’ capabilities brings higher levels of innovation.
Second, horizontal alliances are harder to coordinate especially for knowledge transfer and joint product innovation, but as Mesquita and Lazzarini (2008) state, they result in higher rates of innovation. It is important to note the benefits as well as the risks of horizontal alliances. On the one hand, horizontally linked firms share the same scope of operations and market segments so opportunistic behavior may result in proprietary technology expropriation and imitation (Mesquita & Lazzarini, 2008). On the other hand, being industry peers and having
21 a common stock of knowledge (Sampson, 2007) makes for intense knowledge flows, especially for hard to codify tacit knowledge.
Going further, the nature of the alliance, horizontal or vertical, also has an effect on organizational trust. Organizational trust is in turn connected to the effectiveness of the cooperation between partners (Rindfleisch, 2000). Rindfleisch (2000) concludes that vertical alliances warrant higher levels of organizational trust than horizontal ones. Additionally, while trust influences cooperation in vertical alliances positively, it does not have an effect on horizontal ones (Rindfleisch, 2000). These findings suggest that vertical alliances will have a positive effect on knowledge transfer and innovation output.
It is important to consider trust in the framework of the nature of the alliance for two specific reasons. First, partners that base their interactions on trust are less likely to act opportunistically (Doney & Cannon, 1997; Ganesan & Hess, 1997). Opportunistic behavior is a main concern for born globals from emerging economies. As discussed in the previous sections, born globals’ partners are usually larger MNEs with greater experience, seeking intangible assets like skills and competencies they lack. Consequently, they would have the upper hand in a race to learn compared to their small and inexperienced partners. Second, Rindfleisch (2000) links trust to resource dependency among partners, which differs for horizontal and vertical alliances. Vertically connected partners experience greater interdependence because each provides complementary resources (Gulati, 1995), which requires a greater level of trust and results in more effective collaboration and integration between partners, especially knowledge sharing as noted by Mowery et al. (1996). In the case of born globals from emerging economies, alliance resource dependency is determined also by their asset specificity. As discussed in the literature review section, born globals are characterized by intangible resources like strong knowledge capabilities, inherent innovativeness, superior entrepreneurial orientation and proactivity (Cavusgil & Knight, 2015;
22 Knight & Cavusgil, 2004) and this is what they can contribute to an alliance. Further, born globals coming from emerging economies have developed the additional capabilities of operating in high institutional and market uncertainty (Luo & Tung, 2007). By contrast, traditional MNEs hold tangible assets like financial and human resources besides accumulated experiential knowledge. Having said that, the asset specificity of both partner firms, born globals and traditional MNEs, suggest opportunities for effective collaboration in both forms of alliance, horizontal and vertical. However, according to Rindfleisch’s article (2000) participants in horizontal alliances are less trusting thus weakening the cooperation between them. The author views vertical alliances as ones of “low competitive intensity” (Rindfleisch, 2000, p. 91) which foster high levels of trust and cooperation and low expectations of opportunistic behavior.
Trust is an especially important consideration when choosing a mode of alliance for companies from emerging economies. Taking into account their weak institutional environment and risk of contract enforcement failure (Luo & Tung, 2007), a high level of trust and interdependence between the born global and their partner, coming from a developed or an emerging economy alike, can act as a mitigating factor to institutional failure.
H1: For born globals from emerging economies, horizontal alliances will have a stronger positive influence on innovation output than vertical alliances
MNEs from developed countries leverage their accumulated technological, managerial, marketing and financial strengths to be able to successfully enter foreign markets (Guillen & Garcia-Canal, 2009). By contrast, they fall behind in adaptability and speed of internationalization, which makes their overall competences and capabilities rather complementary to these of a born global. As previously pointed out, born globals generally lack tangible resources but excel in innovation and organizational adaptability (Guillen & Garcia-Canal, 2009). Since companies from emerging economies and companies from
23 developed economies possess heterogeneous tangible and intangible resources, I expect a relationship of complementarity, rather than reinforcement, to have a stronger positive impact on alliance innovativeness. Looking at country of origin from the point of view of Sampson’s study (2007), connecting partners’ diversity and high levels of innovation in alliances. Mainstream IB theory assumes that firms domiciled in the same country share common features at the national level. For example, firms face a common technological and knowledge base, as shown by indices like the World Competitiveness Index (World Economic Forum, 2017) and The Global Innovation Index (GII) (WIPO, Cornell University, & INSEAD, 2017). The fact that the GII takes into account country factors like institutions, infrastructure and market sophistication shows that company innovation cannot be separated from country development levels. Further, IB literature regards such country-level characteristics as integral parts of country choice selection when companies internationalize. For instance, Ghemawat includes administrative and economic distance factors in his CAGE framework (2007), while Berry, Guillén, & Zhou (2010) list economic, financial, political, administrative and knowledge dimensions of distance within their institutional approach. Using such tools to evaluate partner home country compatibility can be useful for an initial screening when making the partner choice decision especially with innovation capabilities in mind. Having in mind that a certain level of diversity is best when aiming for improved innovation capabilities, a company with a home country background supporting such capabilities should have a stronger impact on the alliance innovation performance than a company domiciled in an emerging economy, having the same or lower level of capabilities as the focal born global.
H2: For a born global from an emerging economy, a partner firm coming from a developed country will have a stronger positive influence on alliance innovation output than a partner firm coming from an emerging economy.
24 When considering the effects of partner country origin on alliance form choice, trust between partners is an essential consideration. Research connects the national origin of firms to strategic decision making, including alliance form because of the varying degrees of uncertainty different national origin brings to interfirm linkages (Mayrhofer, 2004). The level of commitment to the JV influences the alliance form choice (horizontal or vertical) as well as partner choice. Moreover, as discussed in the previous sections, organizational trust is an important factor that affects horizontal and vertical alliances differently, with horizontal alliances engendering higher levels of trust (Rindfleisch, 2000). This means that horizontal alliances require greater commitment from both partner firms. IB literature states that high levels of uncertainty result in firms choosing low levels of commitment in order to minimize risk (Johanson & Vahlne, 1977). According to the Uppsala model uncertainty and risk stem from distance, and also internationalizing companies keep commitment to new markets and partnerships low when uncertainty is high (Johanson & Vahlne, 1977). This is why I consider born globals’ partner country of origin an important factor when choosing the form of JV. Academic input on measuring distance is abundant but contradicting (Berry et al., 2010; Hofstede, 1983; Zaheer, Schomaker, & Nachum, 2012) and is not within this study’s scope. However, distance explains the need of this research’s theoretical model to distinguish between traditional MNEs from developed countries and traditional MNEs from emerging economies, more specifically, economic and administrative distance (Berry et al., 2010; Pankaj Ghemawat, 2007). Taking into account the concept of uncertainty and the desired level of commitment it has on JV form choice, I can conclude that in the face of economic distance, firms are more likely to form a joint venture partnership that allows for a low level of commitment. In other words, when both JV partners come from relatively equally developed countries (both the born global and the traditional MNE are from emerging economies), they would be more likely to form a higher trust partnership i.e. a horizontal alliance. By contrast, when the born global
25 coming from an emerging economy has a traditional MNE from a developed economy for a partner, it will be more likely to form a vertical alliance. Going further, commitment and trust in an alliance are critical for successful cooperation especially in terms of knowledge exchange and innovation creation (Wadin, Ahlgren, & Bengtsson, 2017). Regardless of the type of
knowledge in question, codified or tacit, partners’ willingness to cooperate and maintain transparency instead of engaging in a race to learn is vital for the success of the joint venture.
H3: For a born global from an emerging economy, a developed country partner will have a stronger positive moderating effect on the relationship between horizontal vs. vertical alliance and alliance innovation output than a partner firm from an emerging economy.
6. Method
This research is exploratory in nature and its goal is to establish a causal relationship between partner companies and a positive or negative degree of novelty of born globals product portfolio.
6.1. Data sample
To cover as wider and as heterogeneous data sample as possible, I use secondary data, which will also provide a high degree of generalizability of the findings. The final sample has a cross-sectional structure and consists of international JV deals where one of the partners is a born global firm from an emerging economy and the other is a traditional MNEs. Traditional MNEs are further classified according to their country of origin into MNEs domiciled in emerging economies and MNEs domiciled in developed economies. The sample of countries covers all international JV deals available in the database without being geographically limited. The sources I use for data collection are two: Zephyr and Orbis. First, Zephyr offers comprehensive information on international JVs worldwide and systematized data on private companies, necessary for moderator and control variables, and easily accessible. Second, Orbis
26 provides data about every patent associated with a company in the database, which I use in the form of patent counts in order to operationalize the concept of firm innovation.
The data sample includes international JVs founded between 2000 and 2015. After screening and cleaning the sample size from unusable observations due to missing data, 220 observations remained. In this dataset, born globals are identified as such based on two criteria. First, their age at the time the joint venture deal is completed, up to 8 years (McDougall, Shane, & Oviatt, 1994). Unfortunately, this will exclude a number of companies that have been internationalizing on their own before stepping into a joint venture, but neither Orbis, nor Zephyr provides a better way to identify born globals. Second, as discussed in the literature review part, born globals are a type of SME. Since Orbis provides company categorization (very large company, large company, sized company, small company), only medium-sized and small companies are considered as born globals, given they also fulfill the age limit. Tables 1 and 2 provide an overview of the companies comprising the JVs in the dataset in terms of industry affiliation and country of origin.
27 Table 1: Industrial distribution of the MNEs in the dataset
Industry Observations
Manufacturing 93
Financial and insurance activities 32
Wholesale and retail trade; repair of motor vehicles and motorcycles 21
Mining and quarrying 19
Information and communication 14
Transportation and storage 8
Accommodation and food service activities 7
Construction 7
Electricity, gas, steam and air conditioning supply 6
Agriculture, forestry and fishing 4
Professional, scientific and technical activities 3 Administrative and support service activities 2
Human health and social work activities 2
Arts, entertainment and recreation 1
Water supply; sewerage, waste management and remediation activities 1
28 Table 2: Geographical distribution of the companies in the dataset
MNE country of origin Observations MNE country of origin Observations
Japan 37 Sweden 2
Germany 21 Chile 2
Singapore 20 Bermuda 2
The United Kingdom 18 Taiwan 2
France 18 Austria 2
The United States 17 Korea, Rep. 2
Russian Federation 9 Mauritius 2
Switzerland 8 The Philippines 1
Greece 8 Curacao 1
Australia 5 Cyprus 1
Malaysia 4 Egypt, Arab Rep. 1
The Netherlands 4 Botswana 1
Thailand 4 Brazil 1
India 3 Denmark 1
Finland 3 Canada 1
South Africa 3 Colombia 1
Israel 2 Indonesia 1
Romania 2 Hong Kong 1
Belgium 2 Kuwait 1
Slovenia 2 Spain 1
Italy 2 Latvia 1
Born global country of
origin Observations
Born global country of
origin Observations
China 37 Indonesia 3
India 31 Panama 3
Thailand 15 Peru 3
Russian Federation 15 Turkey 2
South Africa 13 Ghana 2
Malaysia 12 Morocco 2
Myanmar 11 Zambia 1
Bulgaria 10 Tanzania 1
Kazakhstan 8 The Philippines 1
Romania 7 Namibia 1 Mexico 6 Kenya 1 Belarus 5 Rwanda 1 Vietnam 5 Uganda 1 Colombia 4 Serbia 1 Ukraine 4 Jordan 1 Brazil 4 Nigeria 1 Croatia 3 Algeria 1 Mauritius 3 Mozambique 1
29 6.2. Variables
6.2.1. Dependent variable
The dependent variable of this study is the patent counts of granted patents and patent applications that occurred after the JV formation. Pakes & Griliches (1984) identify patent counts as a viable way to study firm-level knowledge and innovation, especially compared to RnD expenditure, which is a common alternative approach. I create the dependent variable called JVpat, which is a count variable, taking the values of the JVs count of granted patents and patent applications since their formation. According to Ernst (2001), patent applications and granted patents alike impact company performance, with a stronger influence of the latter. Since the current study measures innovation quantitatively, with count data rather than quality of patents, excluding patent applications would not improve the accuracy of the findings. Moreover, patent applications along with granted patents indicate knowledge exchange and co-creation between within the JV just as much as granted patents. The data for the patents of the JVs in the dataset is sourced from Orbis and then added to the JV data set, exported from Zephyr. Both Orbis and Zephyr are Bureau Van Dijk databases which makes it possible to combine company level data for the JVs from Zephyr with patent data for each JV from Orbis. To do so, I use the unique Bureau Van Dijk identification number every company in both databases is associated with. The export from Orbis contains each JV’s granted patents and patents applications by year of submission ranging from the formation of the JV up to 2017.
6.2.2. Independent variables
Depending on the collected data, characteristics of the alliance partner that I expect to influence the relationship differently can be further examined. For example, Ellis and Mayer (Ellis & Mayer, 2009) suggest that horizontal partnerships benefit NPD and collaboration more than traditional vertical integration.
30
Horizontal vs. vertical alliance
The first independent variable in the study is JV form, horizontal or vertical alliance. In their research on strategic alliances Chan, Kensinger, Keown, & Martin (1997) use United States’ Standard Industrial Classification (SIC) codes to distinguish between horizontal and vertical alliances. If the alliance partners share the same first three digits of their SIC codes, then the alliance is a horizontal one, if the first three digits of their SIC codes are different, then the partners are vertically connected. The first three digits of a SIC code indicate industry group. Furthermore, in the European Commission industry classification NACE Rev.2 the four-digit codes show that a company belongs to an industry class, which corresponds to industry group in the SIC classification (Eurostat, 2008). Due to information availability reasons, in this research I will use the NACE Rev.2 industry classification, relying on Chan et al.’s approach (Chan et al., 1997). To distinguish between horizontal JVs and vertical JVs, I define the dummy variable JVform, which takes the value of one (1) when the JV is horizontal and of zero (0) otherwise, in this case when the JV is vertical.
Partner country of origin
The second independent variable is Partner country of origin. It is a categorical variable, for the traditional MNEs, which are part of the observed international JVs alongside the born globals. The traditional MNEs in the sample are classified according to their country of origin into MNEs from emerging economies and MNEs from developed economies. Zephyr provides a country ISO code denoting country of origin for each company. In addition to that, I use The World Bank country classification by income level (The World Bank, 2017) to distinguish between developed and emerging. The dummy variable MNECCat equals one (1) when the traditional MNE comes from a developed country and zero when it is from an emerging economy.
31
6.2.3. Control variables
Scholars identify industry as a strong determinant of innovation output (Freel, 2003; Oerlemans, Meeus, & Boekema, 1998; Pavitt, 1984). In their respective papers, both Freel (2003) and Oerlemans, Meeus, and Boekema (1998) stress on the importance of not only internal but also external factors on innovation output. The authors reason that outside networks and intra-company linkages vary by industry and provide different opportunities for absorbing and applying knowledge. In his 1984 article Pavitt also looks at the relationship between firm size and innovation (Pavitt, 1984). However, the author notes that the causal relationship is unclear since large firms can enable innovation but innovation can also pre-date size and cause firm growth. The control variable for industry in this study is MNEInd and it denotes the industry section of the traditional MNE according to the NACE Rev.2 industry classification. Since MNE industry is multi-categorical, 15 dummies for each industry section are created. Orbis provides the four-digit NACE Rev.2 industry code for both partners of each observation of the dataset. The first two digits of the code indicate the broad industry section to which the company belongs (Eurostat, 2008). The partner MNEs of the observed joint ventures come from 15 out of the total 21 industry sections.
Other authors also consider firm size as a determinant of innovation (Sourtaris, 1999; Vega-Jurado, Gutiérrez-Gracia, Fernández-de-Lucio, & Manjarrés-Henríquez, 2008). Orbis provides the category of each partner of the JVs in the data sample, based on number of employees and operating revenue. Therefore, to control for company size I use the ordinal variable MNEsize taking values from 1 to 4, where 1 is the category of the smallest company size and 4 is the largest. The four levels of company size are: 1 for “small company”, 2 for “medium-sized company”, 3 for “large company” and 4 for “very large company”. Born globals are by definition SMEs.
32 Partner firms to the born globals in this research can also vary in their age, which is connected to the amount of accumulated experience a company has as well as managerial style in terms of decision-making flexibility, risk-taking and business strategy. This research controls for these effects with the MNEage. It denotes the age of the company taken at the year the international JV is formed. It is calculated by subtracting the year the international JV is formed from the traditional MNE’s year of incorporation. Similarly, two more control variables are calculated, BGage and JVage, controlling for the age of the born global and for the age of the JV respectively. MNEage, BGage and JVage are continuous variables taking positive integer values. The three control variables are created using data from Zephyr.
MNE knowledge base measures the levels of company innovation for the MNE before stepping into the joint venture. The operationalization of the existing knowledge base of the traditional MNE at the time of the JV deal is the same as for the independent variable – patent counts measured from the MNE’s date of incorporation up until and including the year of the JV with a born global’s formation. To control for the effect of knowledge base, the variable MNEpat is defined. It is a scale variable, taking the values of the traditional MNE’s count of granted patents and patent applications from its incorporation up to and including the year it steps into a JV with a born global. The data needed for MNEpat is sources from Orbis.
Similar to the independent variable MNECCat, I define a control variable BGCCat that controls for the heterogeneity between emerging economies. Following the World Bank’s (The World Bank, 2017) country classification by income level that I used to distinguish between developed and emerging countries, emerging economies are divided into three subgroups: low income, lower middle income and upper middle income. This variable controls for nuances in the economic distance within the group of emerging economies and complements hypothesis 2 by adding a second dimension to the distance measured between born globals and traditional MNEs in a JV. BGCCat is an ordinal variable that takes the values of (1) for a low-income
33 economy, (2) for lower middle income and (3) for upper middle income. For this control Orbis provides data for the country of origin, while and The World Bank provides country classification by income level.
DealYear is a variable that aims to control for exogenous factors like the economic cycle (e.g. the financial crisis of 2007-2008) or other chance occurrences. The dataset contains JVs forms between 2000 and 2015 so dummy variables for each year are created. The incorporation year for each international JV is sourced from Zephyr.
An overview of the variables discussed in this section is given in Table 3:
Table 3: Variables definitions and coding
Name Definition Coding Sources
JVpat Post-deal patent counts (patent count for the MNE after forming the JV)
Scale variable Orbis
JVform JV form Dummy variable: 1=Horizontal JV 0=Vertical JV
Zephyr/ NACE Rev.2
MNECCat MNE country category Dummy variable:
1=MNE from developed economy 0=MNE from emerging economy
Zephyr/ The World Bank
BGCCat Born global country category
Ordinal variable: 1=Low income
2=Lower middle income 3=Upper middle income
Zephyr/ The World Bank
MNEage MNE age Scale variable Zephyr
BGage Born global age Scale variable Zephyr
JVage JV age Scale variable Zephyr
MNEsize MNE size Ordinal variable:
1=MNE is a small company 2=MNE is a medium sized company
3=MNE is a large company 4=MNE is a very large company
Orbis
MNEpat MNE knowledge base Scale variable Orbis MNEInd MNE industry I distinguish between 15 industry
sections.
Zephyr/ NACE Rev.2
DealYear Year of JV deal completion Dummy variables for each year a JV was formed ranging from 2000 to 2015
34 7. Statistical analysis
Running Zscore frequencies identified 10 observations as outliers based on JVpat, MNEage and BGage, which were excluded from the analysis. Table 4 shows the dependent variable, independent variables and control variables in this study and their minimums, maximums, means and standard deviations. The table indicates there are no missing values so analysis can be run with all 210 observation of the final sample. The values for the dummy variable JVform have a mean of 0.34 so the majority of observations are of vertical joint ventures. The mean of the other independent variable indicates that most of the joint ventures in the sample are between a born global from an emerging economy and a traditional MNE from a developed economy. JVpat, which is the operationalization for the JVs’ innovation output, and MNEpat, which measures the knowledge base of the traditional MNE partner at the time of the JV deal, both have minimums of 0, which is expected given they are count variables and can take values of positive integers. The mean and max of MNEsize (mean=3.43, max=4) indicate that most traditional MNEs part of the sample are large or very large, which is in line with the theory. The results for standard deviation and mean for MNEage and BGage fall within expectations. Values for MNE age (Std. Dev. = 34.989, mean = 35.00) show that the MNEs in the dataset vary in their age but can all be considered long-term market players. Overall, Table 4 does not identify coding problems.
Table 4 also contains a correlation matrix. The correlation coefficient between the independent variable MNECCat and the dependent variable JVpat is positive and significant (r=0.157, p=0.012), which suggests that MNEs coming from developed countries are positively connected to a higher number of patents. As for the second independent variable, the correlation coefficient for JVform (r=-0.031, p=0.327) is negative but not statistically significant. The correlation between the independent variables MNECCat and JVform is low (r=0.018, p=0.398) and not statistically significant, which excludes the possibility of multicollinearity
35 between the two. There is a very high correlation of 0.805 between the dependent variable JVpat and the control variable MNEpat (r=0.805, p=0.00), which is also significantly different from 0 with a p-value smaller than 0.01. MNEpat controls for MNEs’ knowledge base pre-JV formation and is expected to have a strong positive relationship with post-deal patent counts.
36
Table 4: Descriptive statistics and correlations
Variables 1 2 3 4 5 6 7 8 9 1. JVpat 1 2. JVform -0.031 1 3. MNECCat .157* 0.018 1 4. BGCCat 0.005 0.049 0.013 1 5. MNEage .398** -0.067 .132* 0.044 1 6. BGage -0.063 0.031 -.156* 0.016 0.078 1 7. JVage .161** 0.100 0.101 .208** 0.040 .130* 1 8. MNEsize .307** -0.054 0.038 0.035 .386** -0.012 .135* 1 9. MNEpat .805** -0.039 .200** 0.035 .464** -0.053 0.066 .315** 1 Obs. 210 210 210 210 210 210 210 210 210 Mean 828.33 0.34 0.84 2.67 35.00 0.52 7.75 3.43 4737.28 Std. Dev. 3442.918 0.476 0.369 0.500 34.989 0.934 4.214 1.015 43142.512 Min 0 0 0 1 0 0 2 1 0 Max 28884 1 1 3 157 4 17 4 612933
**. Correlation is significant at the 0.01 level (1-tailed).
37 7.1. Regression analysis
The dependent variable of this study is a count type. Its distribution is overdispersed with a variance exceeding the mean (Var=11853681.02, mean=828.33). Considering the type of dependent variable and its distribution, a Negative Binomial Regression is suitable to test the hypotheses (Yang & Berdine, 2015). In order to test the three hypotheses of the proposed theoretical model, four different Negative Binomial Regression models were run. Model 1 includes only the control variables of the analysis, models 2 and 3 introduce the two independent variables, and model 4 – the interaction term between JVform and MNECCat.
Table 5 contains the results of the Negative Binomial Regression for all four models. The Exp (B) shows the incident rate ratios of the predictor variables and the significance coefficient signals whether the predictor has a statistically significant impact on the dependent variable or not. In addition, model fit statistics are included. The values of the Likelihood Ratio Chi-Square rise from model 1 to model 3, meaning that the independent variables of the study add to the explanatory power of the model. However, model 4 exhibits a slightly lower value than model 3.
M1: JVPat= a0 + a1BGCCat+ a2MNEage + a3BGage + a4JVage + a5MNEsize +
a6MNEpat + a7MNEInd+a8DealYear + Ɛ
Model 1 in Table 5 includes only the control variables of the study and as expected has the lowest likelihood Ratio Chi-square of the five. It is interesting to note the results for BGCCat (Exp(B) = 0.731, p=0.191) and BGage (Exp(B)=1.191, p=0.355). A negative relationship between born globals from higher income emerging economies and JV innovation output and a positive relationship between the age of the born global and JV innovation output. However, both are not statistically significant. Another notable control is MNEsize, which is
38 statistically significant across all models and has a large positive effect (Exp(B)= 6.081, p=0.000).
M2: JVPat= a0 + a1JVform +a2BGCCat + a3MNEage + a4BGage + a5JVage +
a6MNEsize + a7MNEpat + a8MNEInd + a9DealYear +Ɛ
Model 2 in Table 5 presents the Negative Binomial Regression that tests Hypothesis 1, which proposes that horizontal JVs will have a stronger positive influence on innovation output than vertical ones for born global companies from emerging economies. The Exp(B) for JVform, when it equals 0, is lower than 1 and significant (Exp(B)= 0.354, p=0.000). The incident rate for JVform=0 is 0.35 times the incident rate for the reference group JVform=1. In other words, vertical JVs have a weaker influence on JVpat than horizontal JVs. The results of model 2 provide evidence supporting Hypothesis 1. In Model 2, the control variables MNEage (Exp(B)=1.020, p=0.000), JVage (Exp(B)=0.038, p=0.000) and MNEsize (Exp(B)=6.504, p=0.000) increase in magnitude compared to Model 1 and remain statistically significant. By contrast, BGCCat (Exp(B)=0.596, p=0.005) decreases.
