Do leading or lagging firms benefit more from
cross-border innovation?
Alexandra Peckham
10622373
30 June 2014
MSc Business Studies: International Management
Master Thesis
Supervisor: Dr. Niccolò Pisani
Second Supervisor: Dr. Ilir Haxhi
Abstract
Previous literature has established that cross-border innovation has a positive influence on a firm’s innovative performance. However, so far authors have not taken into consideration how this increase in innovative performance affects a firm’s financial performance. Furthermore, although existing research recognizes firm and industry heterogeneity in technological capabilities, it does not address how these factors moderate a firm’s ability to benefit from cross-border innovation. Thus, this study examines the relationship between cross-border innovation and firm performance and the differences between leaders and laggards on a firm and an industry-level by analyzing a sample of Spanish manufacturing firms from 2004-2011. The findings reveal a positive and linear relationship between cross-border innovation and firm performance and a moderating effect of firms and industry heterogeneity in technological capabilities.
Key words: Cross-border innovation; knowledge transfer; firm heterogeneity; industry heterogeneity; firm performance
Table of content
1 Introduction ... 1
2 Literature review ... 5
2.1 Benefits and challenges of cross-border innovation ... 5
2.1.1 Benefits of cross-border innovation ... 5
2.1.2 Challenges of cross-border innovation ... 8
2.2 Technologically leading and lagging firms ... 11
2.3 Cross-border innovation and innovation implications ... 14
3 Theoretical framework ... 17
3.1 Cross-border innovation and firm performance ... 17
3.2 Cross-border innovation and firm performance for leading and lagging firms ... 21
3.3 Cross-border innovation and firm performance in leading and lagging industries ... 23
4 Research methods ... 25 4.1 Sample ... 25 4.2 Variables... 27 4.2.1 Dependent variable ... 27 4.2.2 Independent variable ... 28 4.2.3 Moderating variables ... 28 4.2.3 Control variables... 32 4.3 Statistical method ... 34 4.4 Regression results ... 34 5 Discussion ... 44 5.1 Discussion of results... 44 5.2 Academic relevance ... 46 5.3 Managerial implications ... 47
5.4 Limitations and suggestions for future research ... 48
6 Conclusion ... 49
7 Sources ... 52
List of tables and figures
Table 1: Industry breakdown for the initial sample (year = 2004) ... 27
Table 2: Means and medians of RDI on a firm level by industry ... 31
Table 3: Means and medians of RDIIND by industry ... 32
Table 4: Control variable definition ... 33
Table 5: Descriptive statistics and correlations ... 35
Table 6: Results of regression analysis (full sample) ... 36
Table 7: Results of regression analysis (median split by RDIOECD) ... 40
Table 8: Results of regression analysis (median split by RDISPAIN) ... 41
Table 9: Results of regression analysis (median split by RDIIND) ... 43
1 Introduction
It is widely accepted that research and development (R&D) capabilities are an important influencing factor on firm performance (Bausch & Krist, 2007). Such capabilities form a central part of firms’ competitive advantage (Lee & Rugman, 2012), as science has become an important aspect in many industries (Rothaermel and Hess, 2007). The ability to continuously innovate is essential for survival and success in the competitive global environment (Artz et al., 2010). In order to enhance innovative capabilities, firms strategically source knowledge (Kuemmerle, 1997; Gittelman, 2007), increasingly beyond their national borders (Nieto & Rodríguez, 2011).
Such international knowledge seeking can take on different forms. Firms may choose to establish or acquire subsidiaries abroad (Cantwell & Mudambi, 2005; Ruckman, 2005), enter cooperations with suppliers, customers, universities and other partners (Belderbos et al., 2003) or outsource their activities to offshore third parties (Mukherjee et al. 2013). Firms can also access knowledge through informal mechanisms such as common buyers and suppliers and informal meetings (Alcácer & Chung, 2007). Independent of the organizational form adopted, these activities form the basis of cross-border innovations.
When engaging in cross-border innovation and sourcing knowledge from the external environment, multinational enterprises (MNEs) face the issue of gaining additional knowledge while protecting their own know-how from flowing out to the environment, as a loss of valuable knowledge weakens their competitive position (Belderbos et al., 2003; Perri & Andersson, 2014). Firms have different mechanisms to protect their know-how and ensure the appropriation of profits (Cohen et al., 2002; Branstetter, 2001). The exchange and integration of knowledge within the MNE network is paramount for the utilization of the resource. The implementation of an adequate organizational structure and the necessary inter-unit
coordination are linked with significant challenges, risks and ultimately large investments (Penner-Hahn & Shaver, 2005; Kuemmerle, 1997).
Few authors have examined the extent to which cross-border innovation activities influence a firm’s innovativeness (e.g. Penner-Hahn & Shaver, 2005; Lahiri, 2010; Kotabe et al., 2007). Overall, these studies find that a firm’s ability to innovate improves through international knowledge transfer up to a certain threshold at which the coordination of activities becomes too complex to bring advantages. In other words, there may be certain challenges but the overall impact on a firm’s ability to innovate is concluded to be positive.
Although existing studies establish how a firm’s innovativeness is affected by international
innovation activities, none of them address how a firm is ultimately impacted by cross-border innovation. Although significant benefits are apparent, the costs and risks associated with these innovation activities must be considered. Even if a firm can increase profits by utilizing foreign knowledge to improve existing products or create new products, the costs of achieving these activities are potentially larger than the profits themselves. So far, this connection has however not been addressed in the literature. This study aims to fill this gap by raising the question of how cross-border innovation influences a firm’s financial performance.
Apart from analysing the general relationship between cross-border innovation and firm performance, this study also aims at establishing how the relationship differs depending on the position of a firm in its competitive environment. The extent to which a firm engages in cross-border innovation and experiences the benefits and challenges is highly dependent on the characteristics of its home country-industry on the one hand and its own capabilities on the other hand (Chung & Alcácer, 2002). On an industry-level, the motivation behind firms investing abroad lies in the possibility of tapping into knowledge that does not exist in their home industry (Berry, 2006; Salomon & Jin, 2008). For example, firms from technologically
leading industries are already exposed to sophisticated technology at home and have little to learn from knowledge sourcing abroad (Salomon & Jin, 2008). Firms from technologically lagging industries on the other hand, can use cross-border innovation to compensate for the low level of knowledge at home (Lewin et al., 2009). Thus, firms from leading industries are likely to benefit more from cross-border innovations than firms from lagging industries.
On a firm-level, a firm’s technological abilities are influenced by its industry’s characteristics, however, it cannot be assumed that all firms in an industry possess identical capabilities (Smith, 2014). Whether a firm is technologically advanced or not determines the extent to which it can learn from the external environment and make use of the sourced knowledge (Chung & Alcácer, 2002). Firms with ex-ante technological advantages have the absorptive capacity to recognize, assimilate and process valuable external knowledge better than their lagging competitors and can use it strategically to discover new innovations (Cohen & Levinthal, 1990; Bertrand & Mol, 2013). Thus, absorptive capacity and existing technological skills are a prerequisite for successfully undertaking cross-border innovation that not all firms possess (Penner-Hahn & Shaver, 2005). The existing research in this field recognizes differences between firms’ technological capabilities and also their home country’s industry’s characteristics. Nevertheless, the present literature does not address how these differences have a moderating effect on a firm’s ability to benefit from cross-border innovation. It is unclear how a firm’s capabilities influence its ability to utilize foreign knowledge and how a firm’s home country industry’s position as a relative technological leader or laggard determines how much that firm can benefit from seeking knowledge abroad. Thus, this study aims at clarifying how these ex-ante characteristics influence the relationship between cross-border innovation and firm performance.
1. How does cross-border innovation effect a firm’s performance?
2. Do technologically leading or lagging firms learn more from cross-border innovation?
3. Do firms from technologically leading industries learn more from cross-border innovation than for firms from technologically lagging industries?
On the broadest level this study contributes to the literature on internationalization and performance. Previous studies in this field assume that the costs and benefits of internationalization can be assessed for the firm as a whole. This has led to a number of contradicting theories and models. In reality, different activities, such as R&D and sales, have different characteristics, making it difficult to come to a general conclusion about the impact of internationalization (Alcácer, 2006). Some authors have focused on particular firm activities and assessed their effect on firm performance. For example, Golovko and Valentini (2011) concentrate on the impact of exporting. This present study contributes to the internationalization literature by focusing on R&D activities of a firm in particular and assessing how internationalization of innovation activities influences firm performance.
This study also contributes by enhancing the literature on cross-border innovation. The limited existing literature on cross-border innovation focuses on how international R&D activities affect a firms’ innovative performance (e.g. Penner-Hahn & Shaver, 2005; Lahiri, 2010). This study takes these considerations a step further by assessing the ultimate impact cross-border innovation has on a firm. Furthermore, this study expands the literature by considering the importance of industries’ and firms’ heterogeneity in technological capabilities. This study shows that although all firms benefit from cross-border innovation, the extent of the costs and benefits is determined by a firm’s status as a relative technological leader or laggard on the one hand and a firm’s home country industry’s technological positioning on the other hand. It takes the two strands of literature together and seeks to assess how the degree of cross-border
innovation influences a firm’s performance. Based on the existing literature concerning innovative capabilities, the overall correlation is expected to be positive. However, low and high levels of international activity are expected to entail significant costs due to complex coordination, leading to a negative impact on financial performance. Overall this results in a curvilinear, S-shaped relationship.
The remaining thesis is structured as follows. The following section will review the existing literature on motivations, benefits, risks and costs arising from cross-border innovation activities, taking into account the firm and industry heterogeneity in technological capabilities. Thereafter, the theoretical framework will be introduced. Following this, the dataset, statistical methods and results will be presented. The final section includes the results and discussion of the findings. Concluding remarks end the thesis.
2 Literature review
This section discusses the extant research on the potential benefits and challenges of cross-border innovation. Thereafter, the existing literature on cross-cross-border innovation and innovation performance is presented, taking the differences taking the heterogeneity of firms’ and industries’ technological capabilities into consideration.
2.1 Benefits and challenges of cross-border innovation
2.1.1 Benefits of cross-border innovation
Technological knowledge and innovative capabilities are a central part of firms’ competitive advantage and influence a firm’s performance (Lee & Rugman, 2012). In order to enhance such resources and capabilities, firms seek knowledge spillovers from their external environment. Due to the tacit and localized nature of knowledge, geographical proximity is a prerequisite for accessing local knowledge and benefiting from spillovers (Kuemmerle, 1997; Gittelman,
2007). Therefore, firms strategically position their activities in locations in which the potential for valuable spillovers is high, thus collocating with local actors. The choice of location depends highly on competition in the potential host market. Generally, firms will target locations with large knowledge pools and distinct technological advantages compared to their home location (Alcácer & Chung, 2007; Song & Shin, 2008; Sanna-Randaccio & Veuglers, 2007). Many firms go beyond their national borders in search of valuable knowledge and engage in international innovation activities in order to enhance their resources and capabilities (Nieto & Rodríguez, 2011).
The objective of internationalizing innovation activities therefore lies in accessing knowledge and capabilities from other countries (Penner-Hahn & Shaver, 2005). This local know-how can be used to complement the knowledge originating in the MNE’s home country when adapting products to local demands and supporting the local manufacturing sites. Knowledge can also be transferred back to the headquarters in order to enhance the central technological capabilities (Kuemmerle, 1997) or compensate for limited technological know-how in the firm’s home country (Lewin et al., 2009). Thus, benefits from international R&D stem from bilateral knowledge flows between the activities abroad and the MNEs’ headquarters (Sanna-Randaccio & Veugelers, 2007).
Cross-border innovation activities have become significant for firms’ strategy as the global level of technological capabilities has increased and local knowledge clusters have emerged. Certain countries have become important pools for knowledge (Penner-Hahn & Shaver, 2005), harbouring groups of highly skilled personnel (Manning et al., 2008). Cost considerations are not the driving force behind seeking these types of local resources (Di Minin & Bianchi, 2011; Lewin et al., 2009). Instead, firms decentralize their R&D activities to multiple locations in order to access a wider range of knowledge from an increased number of sources. The gained
knowledge may be used in the subsidiary that originally gathered it or distributed in the MNE network and exploited in another site (Lahiri, 2010; Frost & Zhou, 2005).
To enable the transfer of knowledge within the network, appropriate organisational mechanisms must be in place (Penner-Hahn & Shaver, 2005). Here, a firm has a magnitude of possibilities for organising its innovation network. For example, firms may structure their R&D units according to a hub model in which the MNE’s headquarters remain the central unit, allowing control and access to local knowledge (Criscuolo and Narula, 2007). Many multinational firms have implemented this model to facilitate innovation (Manning et al., 2008). Next to such large-scale mechanisms, firms may choose individual-level solutions for internal organisation. For instance, employees from different geographical locations may be brought together to work on a specific R&D project. These projects can range from long-term developments involving a vast amount of resources to short-term assignments including only a few researchers. By integrating individuals into other organizational sites, a common understanding is created which fosters and encourages learning and knowledge exchange and in turn creates the basis for profitable innovations (Frost & Zhou, 2005).
Inflowing knowledge spillovers generally have a positive effect on firm’s performance. Increased innovativeness of a firm leads to higher profitability (Cho & Pucik, 2005) as the collection of valuable knowledge allows for a wider range of innovations with higher likelihood of success (Leiponen & Helfat, 2011). Thus, local knowledge is considered an asset that leads to increased patent output, innovations, new products and ultimately profits (Penner-Hahn & Shaver 2005). However, firms must protect their gained knowledge and innovations from spilling over to competitors (Sanna-Randaccio & Veugelers, 2007), as this would weaken their competitive position (Belderbos et al., 2003; Perri & Andersson, 2014). Protective mechanisms such as patents are used to ensure that firms appropriate the profits resulting from their innovations (Cohen et al., 2002; Branstetter, 2001).
MNEs may explicitly mandate subsidiaries with developing a competence novel to the entire firm, building on the advantages of the host location (Cantwell & Mudambi, 2005). Firms place these highly specialized sites within knowledge clusters in order to access very specific information (Castellani et al., 2013). Headquarters knowledge is transferred to the foreign subsidiary and combined with local knowledge, allowing the MNE to tap into new lines of business and extend its product portfolio (Blomkvist et al., 2010), thus increasing firm performance (Sukpanich & Rugman, 2007). Furthermore, the level of organisational diversity increases with every additional competence-creating site, bringing benefits for the entire MNE and further enhancing firm performance (Cantwell & Mudambi, 2005).
2.1.2 Challenges of cross-border innovation
Merely establishing R&D sites and collaborations abroad is not sufficient for firms to seek the benefits of cross-border innovation. Instead, firms must coordinate and integrate the knowledge within their network, requiring investments bearing challenges and risks (Penner-Hahn & Shaver, 2005).
When entering new foreign markets, MNEs are confronted with challenges arising from their liability of foreignness (Zaheer, 1995). Here, MNEs face additional costs compared to local actors when acquiring information, due to their unfamiliarity with the host environment (Hymer, 1976). In the context of cross-border innovation, these costs arise when an MNE wants to access tacit knowledge and tap into local knowledge networks. These costs can be reduced by hiring local staff that is well-embedded in the local networks (Singh, 2007). However, the use of local staff bears certain challenges and risks. Managers are most concerned with a decrease in quality of their operations and the security of data. Additionally, the loss of managerial control and inefficient internal coordination poses a significant risk to operations
(Lewin & Couto, 2007). It takes experience and time for managers to find efficient strategies to deal with the difficulties associated with offshore talent (Manning et al., 2008).
In order to make use of the foreign innovations, these activities must be integrated into the existing R&D network (Kuemmerle, 1997), ensuring that the knowledge is coordinated and transferred within the MNE across national boundaries (Frost & Zhou, 2005; Castellani et al. 2013). Geographical distance between the home and host country is not the major concern when communicating and transferring knowledge within the network. Instead, institutional differences make the establishment of common codes and languages necessary to simplify the correct interpretation and absorption of information. The instalment of such mechanisms creates significant costs (Castellani et al., 2013). Moreover, the differences in national and organizational cultures of the individual units can hinder communication and even lead to reluctance to share information within the network (Criscuolo & Narula, 2007).
Efficient knowledge transfer within the network is paramount (Castellani et al., 2013), as the innovativeness of a firm depends on its internal processes (Parachuri, 2010), MNEs must develop routines and organizational structures to enable the use of local knowledge not only in the subsidiary that has obtained it but also in sites in other geographic locations (Frost & Zhou, 2005; Castellani et al., 2013). The challenge lies in finding a suitable organisational model for a firm’s specific situation. For example, a hub model with one central and multiple distributed innovation sites is only suitable for MNEs that face homogenous consumer needs. However, as soon as the requirements and demands become more complex, for instance due to the need to locally adapt products, this model reaches its limits (Garbe & Richter, 2009). Furthermore, the communication amongst the single units is limited, as the highly specialized sites usually do not share the common basis of knowledge necessary for integrating units’ information (Criscuolo and Narula, 2007). Cross-national teams of researchers and collaboration amongst employees may have a common knowledge base but bear the risk of inter-unit rivalry that
causes an unwillingness to share knowledge. Furthermore, removing key researchers from their original setting is a costly exercise and can cause a drop in efficiency (Criscuolo & Narula, 2007).
The choice and establishment of suitable and efficient coordination and control mechanisms is a costly process entailing setbacks and possible failures, particularly in the initial phase when R&D is shifted from one to multiple locations (Belderbos et al., 2003). All possible challenges and risks cannot be identified beforehand, but are discovered and dealt with along the way (Manning et al., 2008). The difficulty behind finding an adequate organizational structure is the constant tension between the creation and protection of knowledge (Perri & Andersson, 2014; Li et al., 2012). Foreign sites must interact and build a relationship with local actors in order to access their knowledge. This requires effort, time and ultimately investments. MNEs cannot be entirely certain of the value of local knowledge before engaging in these relationships, making the investment significantly risky (Perri & Andersson, 2014). Finding the appropriate balance between knowledge inflows and outflows is particularly difficult in collaborations, cooperation and outsourcing, as firms voluntarily share their knowledge with their partners. Firms must find an organizational structure with which they can benefit from knowledge exchange while protecting themselves from unproductive knowledge leakage, for example by limiting opportunism through formal contracts (Li et al., 2012; Mukherjee et al., 2013; Nieto & Rodríguez, 2011). The MNE must then invest in every additional foreign location in order to ensure that the adequate organisation is in place for the new site to integrate into the network and contribute to the overall knowledge pool (Di Minin & Bianchi, 2011).
Even if suitable organizational mechanisms are found, inertia may prevent MNEs from successfully implementing them. Older and more mature sites may be favoured, while smaller and newer sites may be disadvantaged due to their limited resources (Di Minin & Bianchi,
2011). Moving away from established organizational routines is a difficult and costly process (Bausch & Krist, 2007; Criscuolo & Narula, 2007).
Overall, firms must find adequate mechanisms to integrate external knowledge into their network. MNEs must weigh the protection of their own knowledge against the benefits of potential knowledge inflows and assess the costs and risks related to the necessary investments. Finding the best approach is undoubtedly a learning process. Some MNEs respond to this challenge by incrementally increasing their engagement in foreign R&D activities. Here firms begin with internationalising innovative activities that are less critical to competitiveness and to gain experience and confidence that later leads to further expansion (Li & Kozhikode, 2009). Ultimately, this may result in critical R&D activities being performed close to the home base or in a limited number of subsidiaries considered adequate (Di Minin & Bianchi, 2011).
2.2 Technologically leading and lagging firms
The extent to which a firm experiences the challenges and benefits of cross-border innovation is determined by the characteristics of the firm and the industry it operates in. The implications of firm and industry heterogeneity in technological capabilities are discussed in this sub-section.
Leading and lagging industries
A firm’s home country industry influences its characteristics and strategies (Chung & Alcácer, 2002). Some national industries possess factor endowments that are superior relative to the same industry in other countries, giving the industry a comparative advantage (Porter, 1990). Industries in countries with a comparative advantage are considered leaders in the global comparison, whereas those without a comparative advantage are considered laggards (Smith,
2014). The influence of the home country also partly explains the differences in performance of firms operating in the same industry in different countries (Porter, 1991). Specifically, the position of the industry influences the technological capabilities of the firms within it. Firms from a leading industry are more likely to be equipped with superior technological capabilities and abilities to process knowledge compared to their global peers (Chung & Alcácer, 2002; Smith, 2014).
At the same time, home country industry characteristics determine a firm’s motives for investing abroad and its choice of location (Chung & Alcácer, 2002). Generally, firms from a lagging industry are assumed to invest in countries with a leading industry in order to tap into knowledge that does not exist in their home industry (Berry, 2006; Salomon & Jin, 2008), thus compensating for the limited technological know-how at home (Lewin et al., 2009). On the other hand, firms from leading industries are already exposed to sophisticated technology at home and have little to learn from knowledge sourcing abroad (Salomon & Jin, 2008). However, it is unlikely that a firm has access to all relevant technology within its home industry, leaving some potential for seeking valuable knowledge abroad (Castellani et al., 2013).
Leading and lagging firms
Although the industry influences a firm’s technological abilities, it cannot be assumed that all firms in an industry possess identical capabilities (Smith, 2014). In a global comparison, a firm’s technological capabilities are likely to resemble its home industry’s technological positioning. Nevertheless, due to firm heterogeneity in technological capabilities, leaders and laggards exist in any given industry. Whether a firm is technologically advanced or not determines the extent to which it can learn from the external environment and make use of the sourced knowledge (Chung & Alcácer, 2002). Firms with ex-ante technological advantages
have the absorptive capacity to recognize, assimilate and process valuable external knowledge better than their lagging competitors (Cohen & Levinthal, 1990). Only firms with high levels of absorptive capacity can absorb the external knowledge into their activities (Bertrand & Mol, 2013) and strategically use it to discover new innovations. Thus, absorptive capacity and existing technological skills are a prerequisite for successfully undertaking cross-border innovation (Penner-Hahn & Shaver, 2005). Technologically lagging firms may also invest in cross-border innovation but are less able to effectively make use of the accessed knowledge (Knott, 2008). For instance, they will be less able than their leading peers to communicate and transfer the new knowledge across their internal network (Penner-Hahn & Shaver, 2005).
A firm’s position as a leader or laggard also determines its choice of location. As discussed previously, the choice of location is extremely important as its characteristics define the quantity and quality of knowledge available. A firm’s capabilities also predict the types of sources it will be able to access knowledge from (Alcácer & Chung, 2007), as they determine the extent to which a firm can understand external knowledge (García et al., 2012). Being able to make full use of advanced and sophisticated external information requires an equally advanced level of absorptive capacity which only leading firms possess, thus allowing them to choose from a larger variety of knowledge sources. Technologically lagging firms on the other hand, are limited to a small set of possible sources, because of their inadequate capabilities to understand more sophisticated information (Alcácer & Chung, 2007).
The competition within the host country additionally influences a firm’s cross-border innovation behaviour (Sanna-Randaccio & Veugelers, 2007). Leading firms are more likely to suffer from knowledge outflows when positioning themselves in a location dominated by competitors (Alcácer & Chung, 2007). This may lead to advanced firms not engaging in cross-border innovation or at minimum distancing themselves from highly industrial areas (Singh,
2007; Alcácer & Chung, 2007) in order to protect their competitiveness (Alcácer, 2006). Laggard firms possess less valuable knowledge for competitors to acquire (Cantwell, 2009) and can thus locate close to rivals without endangering their own competitiveness (Singh, 2007).
2.3 Cross-border innovation and innovation implications
Taking the possible benefits, challenges and heterogeneity deliberations of cross-border innovation activities together, authors question the ultimate implications of international transfer of knowledge on a firm’s ability to innovate, the quality of these innovations and whether and to what extent challenges form a barrier that is too high to make such activities worthwhile (e.g. Lahiri, 2010; Kotabe et al., 2007; Song & Shin, 2008).
The extent to which R&D activities are distributed geographically influences the quality of the resulting innovations. Innovative quality increases as the MNE makes use of the knowledge sourced in the host locations up to a certain threshold at which the difficulty of coordination reaches a level at which the costs exceed the benefits, hence the quality of innovations is negatively affected. As suggested in extant research, the relationship can be expressed in the shape of an inverted U, in which firms limit the number of foreign R&D subsidiaries in order to keep the coordination at a manageable level while still benefiting from the host countries’ knowledge (Lahiri, 2010; Kotabe et al., 2007).
Further studies have not considered the number of foreign innovation sites but rather focused on the degree of innovation within the subsidiaries and the effect on innovative performance, independent of the amount of subsidiaries. These studies include considerations of the above mentioned heterogeneity amongst firms. As technological capabilities increase, so does the
absorptive capacity for local knowledge from the host country. However, after a certain level of capabilities has been reached, firms have established routines and source less knowledge externally (Song & Shin, 2008). Advanced technological abilities are important as they are a prerequisite for making use of local knowledge and creating innovation from it. If firms can build on existing mechanisms for knowledge transfer, internationalizing innovation activities will lead to increased patent output (Penner-Hahn & Shaver, 2005). Innovation performance will also increase if innovation activities are performed abroad together with a local affiliate, as firms can then benefit from increased direct positive knowledge spillovers and reduced opportunism (Nieto & Rodríguez, 2011).
The few studies covering the topic agree that internationalizing innovation activities up to a certain level improves the innovativeness of a firm, although too much internationalization harms innovation, implying an inverted U-shaped relationship. Most of these studies use a count of patents or patent citations to measure the innovation performance of firms, thus illustrating the increase in quality and quantity of an MNE’s innovation activities (e.g. Kotabe et al., 2007; Lahiri, 2010; Penner-Shaver & Hahn, 2005).
The extent of a firm’s innovativeness has a significant influence on its overall performance. Firms engage in cross-border innovations in order to realise new innovations that will lead to higher returns (Penner-Hahn & Shaver, 2005). Specifically, innovations are expected to meet the new needs of existing customers or the requirements of new customers, potentially in more advanced international markets (Smith, 2014). The ability of firms to continuously innovate and improve products is an important determinant of overall performance as product life cycles are becoming shorter and competition is increasing (Artz et al., 2010; Love & Ganotakis, 2013).
Solely increasing R&D investment does however not result in a raise in firm performance. First and foremost, a firm must possess the resources and capabilities necessary for bringing
innovations to market in order to derive profits from them. Furthermore, it is important to consider the costs a firm incurs for integrating and coordinating the knowledge across its international network, as depicted in detail in the previous section. Particularly at a low and at a high level of cross-border innovation, firms face substantial costs. A firm with a low degree of cross-border innovation will likely be engaging in costly and uncertain trial and error investments in order to find suitable mechanisms for handling its international innovation network (Belderbos et al., 2003; Perri & Andersson, 2014), complicated by cultural and organizational differences (Criscuolo & Narula, 2007). Likewise, a high degree of cross-border innovation results in high costs as firms must coordinate a growing network, expanding in terms of internationality and complexity (Criscuolo & Narula, 2007; Sukpanich & Rugman, 2007), ensuring that each organizational unit is adequately integrated (Castellani et al., 2013).
The existing literature deliberates about cross-border innovation and how a firm can benefit from increasing its engagement in these activities, while discussing the costs and risks that occur. So far, authors have focused on the impact on a firm’s innovativeness rather than how a firm’s performance as a whole is affected. However, as depicted above, an increase in innovativeness leads to significant costs and risks that are not necessarily covered by the benefits. Thus, this study examines the extent to which cross-border innovation impacts a firm’s ultimate performance level and how the effect differs depending on the degree of engagement in the international innovation activities. Considering the difficulties firms face in the early stages of internationalization due to learning, coordination and organisational inertia, firms will have to pass a certain degree until the benefits of internationalizing innovation activities outweigh the costs and the overall impact on firm performance is positive. However, if foreign activities increase beyond a certain threshold, costs are expected to exceed the benefits once again, negatively impacting firm performance, resulting in an overall S-shaped relationship. This expectation will be reasoned in more detail in the following chapter.
3 Theoretical framework
3.1 Cross-border innovation and firm performance
Weighing the benefits and costs against one another, the relationship between cross-border innovation and firm performance is expected to form an S-shape, as depicted in figure 1.
Figure 1: Cross-border innovation and firm performance
Phase I: Negative slope at low levels of cross-border innovation
A low degree of cross-border innovation indicates that a firm is at the beginning of the learning process of handling international innovation networks. In this stage, firms engage in trial and error investments with high costs and uncertain outcomes (Belderbos et al., 2003; Perri & Andersson, 2014). Costs arise from firms coming to terms with the liability of foreignness (Zaheer, 1995) and cultural and organizational differences within the network (Criscuolo & Narula, 2007). As many firms do not follow an R&D internationalization pattern that begins with culturally close countries with subsequent expansion to more distant countries (Ambos, 2005), these differences cause a significant cost factor.
Due to the limited experience of firms in this stage, their ability to absorb knowledge from their host environment is limited (Song & Shin, 2008). Furthermore, many firms imitate leading firms and follow them into popular offshoring locations which possess a valuable knowledge pool but are characterized by a high employee turnover (Manning et al., 2008). A high turnover rate increases the difficulty of enforcing organizational mechanisms (Lewin & Couto, 2007), making the learning process more complex and costly. If efficient organizational mechanisms are not in place, firms have difficulty integrating the foreign knowledge into their network. The complexity of increased knowledge calls for restructuring that may be hindered by organisational inertia (Criscuolo & Narula, 2007). Single R&D projects abroad may go unnoticed by the headquarters. Furthermore, the distance between the remote site and the intellectual property management of the headquarters will lead to the quality of resulting patents suffering, endangering the protection of knowledge towards competitors (Di Minin & Bianchi, 2011). Therefore, in this stage, the firm’s international innovation activities attend to projects that are not highly influential to the overall competitiveness (Li & Kozhikode, 2009).
Overall, low levels of cross-border innovation indicate that such foreign R&D activities are not considered the norm and full integration into the MNE network is not taking place. Firms bear the high costs of R&D operations and learning, however, limited experience, caution and resistance interfere with a full integration and use of the obtained knowledge to improve the firm’s competitive position and ultimately performance. Thus, the costs resulting from operating, coordinating and setting up facilities are expected to outweigh the benefits.
Phase II: Positive slope at medium levels of cross-border innovation
At a medium level of cross-border innovation, these foreign innovation activities are more integrated in the MNE network. As experience increases, so does a firm’s ability to source and
absorb a wide range of external knowledge efficiently and find adequate organizational mechanisms to integrate and coordinate this valuable knowledge within its network (Criscuolo & Narula, 2007; Lahiri, 2010; Song & Shin, 2008). This strengthens the firm’s technological capabilities and results in higher levels of cross-border innovation (Penner-Hahn & Shaver, 2005).
The increased experience in cross-border innovation activities reduces the costs and risks associated with operations. For instance, firms gain experience regarding the handling of offshore staff and learn effective mechanisms to hire, improve and retain this important knowledge input. This in turn enhances the flow of knowledge and increases efficiency (Manning et al., 2008; Lewin et al., 2009). Furthermore, firms gain the ability to detect locations most suitable for R&D activities, i.e. those with an adequate knowledge pool and intellectual property protection, and establish centres of excellence within these clusters (Di Minin & Bianchi, 2011; Cantwell & Mudambi 2005). These sites will be assigned with innovation tasks closely related to the firm’s overall competitiveness (Di Minin & Bianchi, 2011; Li & Kozhikode, 2009). The high level of expertise in these centres will allow firms to access new technologies and enhance their product portfolio as capabilities of separate organizational entities are combined through inter-unit learning and new skills are created (Blomkvist et al., 2010; Lahiri, 2010). The resulting increased product diversity enhances firms’ performance (Sukpanich & Rugman, 2012). Furthermore, the improved familiarity with the local environment enables firms to adapt their products to local needs and stimulate sales growth within that particular country (Lewin et al., 2009).
Overall, at a medium level of cross-border innovation, the costs of coordinating activities within the network remain. Firms must continuously enhance mechanisms of information-sharing to make use of created knowledge. However, as experience increases, the costs related
to learning decrease. The benefits a firm reaps from the increased patent output, strengthened technological capabilities and resulting innovations are expected to exceed these costs and lead to an increase in overall performance.
Through the enhancement of coordination and communication, the MNE as a whole improves its ability to source knowledge from its competitors (Sanna-Randaccio & Veugelers, 2007).
Phase III: Negative slope at high levels of cross-border innovation
A growth in cross-border innovation activities results from firms establishing more foreign R&D sites or the expansion of activities within existing sites. Either way, the impact on firm performance is expected to be negative after a certain threshold is reached.
If firms expand their activities further into new countries, the further benefits gained will not be large enough to outweigh the additional costs incurred. Countries with low cultural and trade barriers will have already been entered, leaving only countries with higher barriers and thus increased cost. If the new countries and previously entered countries are very similar, they will not add any value in order to stimulate additional sales (Qian et al., 2008). Furthermore, if further units are added to a network of established sites, this can cause rivalry, hindering knowledge exchange (Criscuolo & Narula, 2007). Nevertheless, the firm incurs the costs of setting up organizational units and integrating them into the network (Castellani et al., 2013).
If a firm’s centres of excellence continue to recombine knowledge and expand the product portfolio, this will eventually harm performance. As product diversity increases beyond a certain threshold, the costs of coordination exceed the benefits, causing a negative impact on performance (Sukpanich & Rugman, 2007).
As the degree of cross-border innovation exceeds a certain threshold, firms can simply not process additional information and knowledge effectively. An overload of information and knowledge in form of an increased amount of patents can harm the quality of the accompanying innovations (Lahiri, 2010; Kotabe et al., 2007). Firms may unintentionally overinternationalize or strategically choose to do so (Contractor et al., 2003).
Overall, the degree of cross-border innovation reaches a threshold at which the incremental costs of increased complexity and coordination are no longer covered by the additional benefits. Even if innovations increases, the developed knowledge cannot be processed to an extent at which performance increases.
Taking these three stages together, this leads to the following hypothesis.
Hypothesis 1: The relationship between cross-border innovation and firm performance is nonlinear, with a negative slope at low levels of cross-border innovation, a positive slope at medium levels of cross-border innovation and a negative slope at high levels of cross-border innovation.
3.2 Cross-border innovation and firm performance for leading and lagging firms
Firms are heterogeneous in their technological capabilities, skills and resources, affecting their ability to utilize foreign knowledge (García et al., 2012). Technologically leading firms are characterised by a high absorptive capacity, enabling them to recognize valuable knowledge, integrate it into their existing networks (García et al., 2012) and strategically make use of it (Berry, 2006). Through their abilities, they can utilize more advanced knowledge from a large variety of sources (Alcácer & Chung, 2007). Although they are in danger of losing valuable knowledge to their foreign peers, their superior position and experience also brings significant
advantages. For example, leading firms are more able to establish themselves in a foreign network (Cantwell, 2009), are more attractive to potential foreign partners and employees (Penner-Hahn & Shaver, 2005) and can use their experience to build more effective contracts (Vanneste & Puranam, 2010). Furthermore, leading firms have already established mechanisms for communication and knowledge transfer within their network, allowing for the efficient use of sourced knowledge (Castellani et al., 2013; Penner-Hahn & Shaver, 2005) and increase in innovative output (Penner-Hahn & Shaver, 2005).
This prior experience allows leading firms to mitigate some of the costs and risks associated with cross-border innovation (Penner-Hahn & Shaver, 2005). Compared to lagging firms, leading firms can start at a later point in the learning curve and avoid a share of the trial and error investments (Salomon & Jin, 2010). The establishment of new organisational units and their integration in the networks still requires investments but the effort, costs and uncertainty are reduced due to prior experience (Song & Shin, 2008).
Compared to leading firms, lagging firms are not equipped with the experience and advanced capabilities that would facilitate cross-border innovation. Due to their limited technological abilities, these firms have more to learn from international innovative activities, but lack the resources to strategically make use of the sourced knowledge and bring the innovations to market in order to achieve profit (Love & Ganotakis, 2013; Berry, 2006). Lagging firms must establish adequate mechanisms for absorbing knowledge from the external environment, and integrating, coordinating and transferring it internally without being able to build on existing capabilities and experience, thus increasing substantial cost and risk (Berry, 2006). Furthermore, due to low absorptive capacity, lagging firms must limit themselves to less advanced knowledge from a narrower selection of sources (Alcácer & Chung, 2007).
Overall, when engaging in cross-border innovation, technologically leading firms can benefit from a larger range of knowledge while reducing some of the costs involved in these activities. Lagging firms only have a limited choice of knowledge they can seek and must bear a substantial cost to utilize this knowledge and benefit from cross-border innovation. The expected shape of the relationship-curve remains an S-shape, as both categories of firms face the same costs and benefits, albeit to different extents. Taken together, leading firms are therefore expected to achieve a higher level of firm performance compared to lagging firms, when engaging in cross-border innovation, leading to a steeper curve. This expected result is denoted in the following hypothesis:
Hypothesis 2: All else equal, the relationship between cross-border innovation and firm performance is stronger for technologically leading firms than for technologically lagging firms.
3.3 Cross-border innovation and firm performance in leading and lagging industries
Whether a firm’s industry is technologically leading or lagging compared to the same industry in other countries also determines how much a firm can learn from cross-border innovation. Specifically, a firm from a leading industry is already exposed to sophisticated technology. Thus the opportunity for learning outside of the national industry is limited (Salomon & Jin, 2010) and sourcing knowledge locally may be more effective (Smith, 2014). Nevertheless, firms are unlikely to be able to access all relevant technological knowledge at home, even if they originate from a leading industry, allowing for the possibility of learning abroad (Castellani et al., 2013).
Firms stemming from lagging industries on the other hand have a great deal to learn from cross-border innovation. By locating their innovative activities in technologically advanced areas,
they gain access to industry-specific knowledge that does not exist in their home country (Smith, 2014; Salomon & Jin, 2010), thus compensating for the lack of valuable domestic spillovers (Lewin et al., 2009). The exposure to more sophisticated markets in leading industries allows firms to improve their capabilities and skills (Salomon & Jin, 2008) as they can use more advanced peers as their benchmark (Smith, 2014). Thus, through cross-border innovation, firms from lagging industries can gain knowledge in order to catch up with their global peers and improve their competitive position on a local and a global scale (Salomon & Jin, 2008; Smith, 2014).
Taken together, firms operating in industries in which their home country is a technological laggard have a higher potential of improving their innovative performance through engaging in cross-border innovation. As firms from leading industries are already exposed to advanced technology, their possibilities for learning abroad are significantly lower, but still existent. The industry’s global positioning as a leader or laggard, does however not influence the necessary costs for cross-border innovation. Without considering other determinants, firms from lagging industries will thus be able to improve their financial performance more than firms from leading firms. Assuming costs are equal for both types of industries, both firms from leading and lagging industries are expected to reveal an S-shaped relationship between cross-border innovation and firm performance, with firms from leading industries showing a flatter curve than firms from lagging industries. Formally, this is stated in the following hypothesis:
Hypothesis 3: All else equal, the relationship between cross-border innovation and firm performance is stronger for firms from industries in which the home country is a technological laggard.
4 Research methods 4.1 Sample
The analysis is based on a dataset derived from a panel survey of Spanish manufacturing firms. The survey has been carried out by the SEPI Foundation and the Spanish Ministry of Industry since 1990 on a yearly basis. This study covers the period from 2004 to 2011. Each yearly survey covers approximately 1800 firms and thus allows the examination of a large amount of data. It is very detailed, including information on firms’ customers and suppliers, activities, served markets, technological activity etc. It guarantees a high level of reliability and quality. The representativeness is assured as firms are selected according to their size and characteristics matching those of firms in past samples and representing the range of firms in the Spanish manufacturing industry. This ensures the absence of any sample bias. Furthermore, the majority of answers capture objective measurements and do not depend on respondents’ perception, making common method bias and subjectivity less likely. Finally, the value and quality of the dataset is demonstrated by its use in a large number of articles published in leading management journals (e.g. Campa and Guillén, 1999; Cassiman and Golovko, 2011; Fernández and Nieto, 2006).
In addition to the before mentioned advantages, the dataset is also particularly suited for the analysis of this study. Spain provides a suitable empirical setting as it holds an average position within OECD countries in terms of R&D expenditure (OECD, 2012). Data from countries considered technological leaders, such as Japan and the US, or countries with particularly low R&D expenditure, could lead to misleading results. By analyzing a middle-ranked country, the findings should be generalizable to other countries and settings, as assumed by previous studies (e.g. Nieto and Rodríguez, 2011). Furthermore, the manufacturing industry has been the empirical setting of multiple studies (e.g. Ruigrok and Wagner 2003, Lahiri, 2010).
In order to make the differentiation between firm and industry leaders and laggards, further data is collected from the OECD Analytical Business Enterprise Research and Development (ANBERD) database. It includes annual data on R&D expenditure, listed by country and country-industry. Moreover, information on countries’ gross domestic products and producer price indexes is obtained from the OECD database. The data is obtained per year for each year in the sample, i.e. from 2004 to 2011.
Statistics on the sample
In total, the survey covers the years 2004 to 2011. The initial sample in the year 2004 consists of data from 1,374 firms in 20 distinct industries. To ensure consistency with the OECD data, firms in the category “Other manufacturing” were excluded from the sample, sacrificing 31 observations. Furthermore, three industries are consolidated to the industry food products,
beverages and tobacco. This leaves 1,343 firms from 17 unique industries in the initial sample.
Table 1 gives an overview of the amount of firms in the industry in the initial sample year 2004 together with some descriptive statistics.
Over the years, firms drop out of the sample and are replaced with firms of a similar size and from the same industry. Thus, the sample grows to an unbalanced panel dataset with a total of 3,003 unique firms. 370 firms from the industry “other manufacturing” are removed to ensure consistency with the OECD database. Due to the lags incorporated in the regression and missing values, another 454 firms are deleted from the data set. This leaves an unbalanced panel of 2,549 firms and 11,084 firm-year observations.
Table 1: Industry breakdown for the initial sample (year = 2004) Industry Number of firms Percentage of total Employees per firm Sales per firm (in millions) Foreign patents per firm 1. Food products. beverages
and tobacco
186 14% 275.53 84.91 0.02
2. Fabricated metal products 166 12% 140.52 21.98 0.07
3. Textiles and clothing 114 8% 97.64 10.39 0.01
4. Chemicals and pharmaceuticals
90 7% 340.01 131.84 3.52
5. Non-metal mineral products 97 7% 257.26 54.74 0.00
6. Machinery and equipment 88 7% 164.98 30.52 0.14
7. Printing 76 6% 217.36 41.98 0.00
8. Plastic and rubber products 76 6% 325.74 65.24 0.05
9. Furniture 74 6% 104.53 12.80 0.32
10. Electric materials and accessories
63 5% 312.63 67.08 0.48
11. Vehicles and accessories 73 5% 1126.75 502.46 0.22
12. Basic metal products 51 4% 514.06 203.74 0.04
13. Timber 46 3% 144.37 19.49 0.00
14. Paper 46 3% 240.33 57.92 0.00
15. Leather. fur and footwear 31 2% 50.77 5.28 0.16
16. Computer products. electronics and optical
33 2% 378.55 112.19 0.15
17. Other transport equipment 33 2% 539.15 116.32 0.03
Full sample 1343 100% 284.82 83.50 0.32
4.2 Variables
4.2.1 Dependent variable
Following previous studies (e.g. Contractor et al., 2003), return on sales (ROS) proxies for firm performance. ROS is measured as labour costs subtracted from the added value and divided by production and other income. In order to improve the interpretation of results the variable is standardized by subtracting the mean and dividing by the standard deviation (Gelman, 2008).
4.2.2 Independent variable
In order to measure the degree of international innovation activities, the ratio of patent applications abroad to total patent applications is chosen as a proxy for the independent variable
cross-border innovation. Patents are an important protective instrument to ensure the
appropriation of profits related to know-how (Cohen et al., 2002; Branstetter, 2001). Furthermore, patents demonstrate a firm’s ability to create something new (Grilichis, 1990) and enable knowledge exchange as tacit information is codified (Jung & Lee, 2010). Thus, patents can proxy for knowledge gained abroad. As knowledge takes time to flow through the company and convert to profits (Salomon & Jin, 2007), the patent ratio is lagged by one year. Furthermore, the square and cubic term of the independent variable is calculated, in order to test for the hypothesized S-shaped curve.
4.2.3 Moderating variables
To distinguish between leaders and laggards on the industry and firm level, three different research and development indexes (RDIs) are created, based on those computed by Salomon and Jin (2007; 2010). The indexes take a firm’s and its industry’s R&D expenditure into consideration. On a firm-level, R&D expenditure proxies for technological capabilities. For example, the extent of a firm’s absorptive capacity is considered the result of its R&D expenditure (Cohen & Levinthal, 1990). On an industry-level, the R&D expenditure indicates its technological endowments (Chung & Alcácer, 2007). Thus, for both firms and industries technological leaders have a relatively high level of R&D expenditure, whereas technological laggards are characterized by a relatively low level of R&D expenditure.
For the differentiation between leaders and laggards on a firm-level two indexes are created. The first index compares a firm with its peers in the same industry across OECD countries, excluding Spain. The R&D expenditure of each firm i in industry j and year t is scaled with its sales in year t to account for size effects. Similarly, the R&D expenditure of each OECD industry j in country k and year k is deflated with its gross production in industry j in country k and year t, as it measures the market value of finished goods per industry and thus is the best comparison to firms’ sales on a macro level (Salomon & Jin, 2010). The average of all scaled R&D expenditures of the OECD countries in industry j and year t is then subtracted from the scaled R&D expenditure of each firm i in industry j and year t. Information on R&D expenditure and GDP are derived from the OECD and ANBERD database. All values are recorded in purchasing power parity (PPP) US dollars, to ensure consistency and account for year effects.
This calculation is depicted in the following formula for firm i in industry j and year t:
𝑅𝐷𝐼𝑖𝑗𝑡𝑂𝐸𝐶𝐷= 𝑅𝐷𝑖𝑗𝑡 𝑆𝑎𝑙𝑒𝑠𝑖𝑗𝑡 − [∑𝑅𝐷𝑗𝑘𝑡 𝑂𝐸𝐶𝐷 𝐺𝑃𝑗𝑘𝑡𝑂𝐸𝐶𝐷 𝐾 𝑘=1 ] ∗ 1 𝐾 Where,
𝑅𝐷𝐼𝑖𝑗𝑡𝑂𝐸𝐶𝐷: RDI for firm i in industry j at time t
𝑅𝐷𝑖𝑗𝑡: R&D expenditure of firm i in industry j at time t
𝑆𝑎𝑙𝑒𝑠𝑖𝑗𝑡: Total sales of firm i in industry j at time t
𝑅𝐷𝑗𝑘𝑡𝑂𝐸𝐶𝐷: R&D expenditure in industry j from country k at time t
This index compares firms to their corresponding global average. Increasing RDI values indicate that the firm is technologically leading, decreasing values indicate that the firm is a laggard.
A second index compares firms with their industry-peers in Spain. Thus this index runs a comparison on a domestic level. The R&D expenditure of each industry j in Spain at time t and deflated with the gross production of the respective industry j at time t. This value is subtracted from the R&D expenditures of each firm i in industry j at time t, deflated by the firm’s sales at time t. This calculation is depicted in the following formula for firm i in industry j and year t:
𝑅𝐷𝐼𝑖𝑗𝑡𝑆𝑝𝑎𝑖𝑛= 𝑅𝐷𝑖𝑗𝑡 𝑆𝑎𝑙𝑒𝑠𝑖𝑗𝑡
− 𝑅𝐷𝑗𝑡
𝑆𝑝𝑎𝑖𝑛
𝐺𝑃𝑗𝑡𝑆𝑝𝑎𝑖𝑛
Again, increasing RDIs values indicate that the firm is technologically leading compared to its home country’s industry average. Decreasing values indicate that the firm is technologically lagging.
To distinguish between leaders and laggards, firms in the sample are split into two categories according to the median of RDI, at first using the RDIOECD, then the RDISPAIN. Firms with an RDI above the median are considered technological leaders whereas firms with an RDI below the median are considered technological laggards.
Table 2 shows the mean and median of RDIOECD and RDISPAIN calculated on a firm level by industry. Comparing the two RDIs, the means and medians of RDISPAIN are larger than those
for RDIOECD. The firms in the sample can thus generally be considered technologically more advanced compared to their local peers and technologically less advanced in comparison with firms in other OECD countries.
Table 2: Means and medians of RDI on a firm level by industry Industry RDIOECD Mean RDIOECD Median RDISPAIN Mean RDISPAIN Median 1. Food products, beverages and tobacco 0.000 -0.003 0.003 0.000
2. Fabricated metal products -0.001 -0.005 0.004 0.000
3. Textiles and clothing -0.001 -0.006 0.005 0.000
4. Chemicals and pharmaceuticals -0.017 -0.030 0.019 0.006
5. Non-metal mineral products -0.003 -0.006 0.002 0.000
6. Machinery and equipment -0.003 -0.016 0.017 0.003
7. Printing -0.001 -0.002 0.001 0.000
8. Plastic and rubber products -0.004 -0.009 0.007 0.000
9. Furniture 0.002 -0.002 0.002 0.000
10. Electric materials and accessories -0.016 -0.017 0.012 0.001
11. Vehicles and accessories -0.007 -0.016 0.015 0.000
12. Basic metal products 0.000 -0.004 0.004 0.000
13. Timber 0.001 -0.001 0.001 0.000
14. Paper -0.002 -0.003 0.001 0.000
15. Leather, fur and footwear 0.000 -0.005 0.005 0.000
16. Computer products, electronics and optical -0.037 -0.059 0.042 0.019
17. Other transport equipment 0.024 -0.037 0.077 0.000
Full sample -0.004 -0.004 0.013 0.000
The R&D index on the industry level is computed in a similar manner. The technological intensity for each industry in Spain is compared with the average of the OECD countries. The annual R&D expenditures of all countries were derived from the ANBERD database. In order to account for size effects, the R&D expenditures are deflated with the GDP of the particular country, which is taken from the OECD database. Again, all values are recorded in PPP US dollars, to ensure consistency and account for year effects. The deflated average of the OECD countries (excluding Spain) is subtracted from the value for each year and industry. Increasing values indicate that the industry is a technological leader, decreasing values indicate that the industry is a technological laggard. Formally, the following formula depicts the RDI for industry j in year t. 𝑅𝐷𝐼𝑗𝑡𝐼𝑁𝐷 = 𝑅𝐷𝑗𝑡 𝑆𝑝𝑎𝑖𝑛 𝐺𝐷𝑃𝑡𝑆𝑝𝑎𝑖𝑛− [∑ ( 𝑅𝐷𝑗𝑘𝑡𝑂𝐸𝐶𝐷 𝐺𝐷𝑃𝑘𝑡𝑂𝐸𝐶𝐷 ) 𝐾 𝑘=1 ]∗ 1 𝐾
The means and medians of RDIIND are depicted in table 3.
Table 3: Means and medians of RDIIND by industry
Industry
RDIIND Mean
RDIIND Median 1. Food products, beverages and tobacco -0.018 -0.022
2. Fabricated metal products -0.050 -0.050
3. Textiles and clothing -0.002 0.001
4. Chemicals and pharmaceuticals -0.927 -0.949
5. Non-metal mineral products -0.010 -0.016
6. Machinery and equipment -2.537 -2.534
7. Printing -0.005 0.000
8. Plastic and rubber products -0.089 -0.084
9. Furniture -0.003 -0.001
10. Electric materials and accessories -0.588 -0.610
11. Vehicles and accessories -0.725 -0.783
12. Basic metal products -0.139 -0.150
13. Timber -0.011 -0.010
14. Paper -0.047 -0.043
15. Leather, fur and footwear 0.007 0.006
16. Computer products, electronics and optical -0.228 -0.275
17. Other transport equipment 0.085 0.118
Full sample -0.311 -0.036
When taking the average across eight years as done in table 3, the RDIINDs are negative for all but two industry. This indicates that in general Spain is technologically lagging compared to other OECD countries.
4.2.3 Control variables
Following previous studies certain determinants of firm performance are included as control variables, namely, firm age, firm size, degree of foreign ownership, the R&D intensity within the firm and organizational changes of the firm.
The age of a firm may influence its decision to internationalize R&D. For instance, younger firms may be able to adapt to new structures easier, whereas older firms risk organizational inertia (Bausch and Krist, 2007). Firm age is calculated as difference between the founding year of the firm and the year the survey was completed. Firm size proxies for the financial and physical resources a firm has at its disposal. It is measured as the natural logarithm of total employees (Qian et al., 2008). Being part of a multinational or foreign company can influence firms’ willingness to invest abroad and indicates previous international experience as well as learning through foreign ownership. The variable is captured as the degree of foreign capital (Almodóvar and Rugman, 2014). R&D intensity proxies for the knowledge available within the firm. It is calculated as the ratio of total R&D expenditure to total sales (Argyres and Silverman, 2004; Lin et al., 2011). Organizational change in a company, for instance a merger, acquisition or a split, may bring about an abrupt increase in sales (Golovko & Valentini, 2011). To control for this, a dummy variable is included that takes the value of 1 when an organizational change takes place at time t and a value of 0 when no organizational change takes place. To control for general differences amongst industries, the 17 industries are consolidated into 6 categories. 5 dummy variables are included to account for industry heterogeneity. The control variables and their calculation are depicted in table 4.
Table 4: Control variable definition
Control variable Calculation
Firm size Natural logarithm of the number of employees R&D intensity R&D expenses divided by total sales
Firm age Difference between founding year and year of the survey Foreign ownership Percentage of foreign ownership in the firm
Organizational change Dummy variable, 1 = change, 0 = no change Industry 5 dummy variables for the 17 discreet industries
4.3 Statistical method
A hierarchical regression is used to test the relationship between cross-border innovation and firm performance. Due to the panel structure of the data, the possibility of serial correlations of residuals of firms with several observations must be taken into consideration, as they could lead to biased results. Following Salomon & Jin (2007) a dynamic longitudinal model is used which incorporates an autoregressive process that includes a lagged value of the dependent variable in the regression (Alzaid & Al-Osh, 1990). This lagged value takes into account the extent to which firms are heterogeneous, beyond what is captured with the control variables (Greene, 2003). As a one-year lag of the independent variable is used, the dependent variable is also lagged by one year and included in the regression to account for firm-specific effects. To ensure consistency, the lagged dependent variable is also standardized. In order to consider time-specific effects, year dummies are included in the regression.
The hierarchical regression consists of four models. Model 1 is the base model and only contains control variables. In model 2 the independent variable cross-border innovation is added to the regression. As the relationship between cross-border innovation and firm performance is hypothesized as being an S-shape, the square and cubic term of the independent variable are included in model 3 and 4, respectively.
To test hypothesis 1, the regression is performed on the full sample. Thereafter, the sample is split into leaders and laggards according to the RDI medians to test hypotheses 2 and 3 concerning firm and industry heterogeneity in technological capabilities.
4.4 Regression results
Table 5: Descriptive statistics and correlations Mean S.D. 1 2 3 4 5 6 7 8 1. ROS -0.025 1.052 1 2. Age 28.697 20.252 -0.004 1 3. Size 4.184 1.475 0.068*** 0.298*** 1 4. Change 0.038 0.192 0.006 0.03*** 0.144*** 1 5. Foreign ownership 14.989 34.811 0 0.17*** 0.459*** 0.107*** 1 6. R&D intensity 0.808 2.764 -0.025*** 0.071*** 0.157*** 0.031*** 0.052*** 1 7. ROSt-1 0.000 1.000 0.4*** 0.004 0.075*** 0.01 -0.002 -0.003 1 8. Cross-border innovation 0.025 0.137 0.025*** 0.073*** 0.196*** 0.03*** 0.068*** 0.182*** 0.015** 1 *p<0.15, **p<0.1, ***p<0.05
