ࡱ > [ F bjbjUU B ?b?b1 % ~
~
X R @ J J J ^ ^ ^ 8 ^ f$ % " .% .% .% 2 j
[= W@ $ ƣ | - J 2X 1 N 2 2X 2X - .% .% u B , h h h 2X < .% J .% h 2X h h l | ϟ .% @bY F] * n H A 6 :d 6 T ϟ 6 J ϟ $ A L #I 6 h YN , R A A A - - 4h A A A 2X 2X 2X 2X 6 A A A A A A A A A ~
B :
Are innovation and internationalization related? An analysis of European countries
Andrea Filippetti(, Marion Frenz( and Grazia Ietto-Gillies(
Are innovation and internationalization related? An analysis of European countries
Abstract
This paper examines the relationship between countries international profile and their innovation performance using data for 32 European countries. The overall contribution of the paper lies in (a) an in-depth exploration of empirical correlations between innovation and several indicators of internationalization; and (b) the use of theoretical arguments backed up by the literature on why the observed correlations are not spurious but indicative of possible causality. Indicators of internationalization are considered with respect to each country as aggregate, to its technology-intensive industries only, and in relation to the share of its firms reporting international activities. On the basis of the empirical results, and the theoretical arguments presented, the paper suggests that underpinning the association is a virtuous (or vicious) circle: innovative firms are more successful in competing internationally and the exposure to alternative business and innovation contexts leads to innovation.
Keywords: Innovation, Internationalization, European Countries, Innobarometer, Community Innovation Survey
JEL classifications: F20, O19, O30
1. Introduction
The research underlying this paper links innovation to internationalization. It takes as its starting point the innovation performance of countries as measured in the European Innovation Scoreboard (EIS). The EIS is an annual report managed by the European Commission Directorate General Enterprises and Industry and carried out since 2001; it measures and compares the innovation performance of countries using a synthetic composite indicator: the Summary Innovation Index (SII). The latter is based on 29 variables addressing several dimension of a countrys system of innovation (see Appendix I). The EIS 2008 includes innovation indicators and trend analyses for the EU27 Member States as well as for Croatia, Turkey, Iceland, Norway and Switzerland ADDIN EN.CITE European Commission200959259227European Commission,European Innovation Scoreboard 2008. Comparative analysis of innovation performance2009BrusselsEuropean Commission, DG Enterprise(European Commission, 2009a); this is the group of 32 countries considered in this paper. The boundaries of the research in terms of content, structure and number of countries included in the analysis are largely set by the EIS.
Considerable progress has been made to reveal cross-country patterns of innovation performance, made possible largely through the coordinated efforts by (European) governments to collect relevant data through the Community Innovation Surveys (CISs). However, less progress has been made towards systematically capturing the global embeddedness of the activities of countries and linking these to the innovation performance of countries.
This paper examines the association between the degree of international embeddedness of business activities and innovation performance across the above 32 countries. The relationship between innovation (from EIS data) and internationalization is considered for the national aggregate, i.e. the country as a whole is the unit of analysis.
The overall contribution of the paper lies in the following: (a) the in-depth exploration of empirical correlations between innovation and several indicators of internationalization for 32 countries; and (b) the use of theoretical arguments backed up by the literature to highlight the existence of causal mechanisms. This supports our thesis that the correlations are not spurious but indicative of causation.
Specific contributions of the paper are the following. Firstly, the identification of three levels of analyses for internationalization: countries internationalisation with respect to all industries, here considered as level A; countries internationalization with respect to innovation-intensive industries referred to as level B; countries internationalization measured as the proportion of enterprises that reported international activities, as level (C). Secondly, the identification of sets of variables within each level, the transformation of variables into indicators following a procedure consistent with that of the EIS and the calculation of Summary Globalisation Indices (SGIs) for each internationalization subset (A, B and C). Thirdly, the calculation of partial correlation coefficients between indicators of innovation (taken from the EIS) and indicators of internationalization at the three levels of analysis. Finally, the analysis of the results in the context of the relevant theoretical discussions on the relationship between innovation and internationalization.
Thus the paper provides a systematic analysis of patterns of association of the international profile with countries innovation performance via partial correlation coefficients controlling for a set of country characteristics. The data used in this paper comes from official national statistics and from Eurostat. From a methodological point of view, we are aware that there are limitations on drawing implications from correlations. The correlations: (a) could be spurious; and (b) even if not spurious they do not give us indications of the direction of causality. We are inclined to rule out (a) because the association between innovation and internationalization is well founded; there are plausible theoretical explanations providing mechanisms for a causal link between these two economic elements. These theoretical explanations will be explored further in the next section, as will be the issue of the direction of causality (point b above). While the paper cannot reach definite conclusions on the direction of causality, the results support the view that causality exists and that they can therefore be used as springboard for further research.
The paper is structured as follows. Section 2 provides the theoretical context and discusses the literature on the relationship between innovation and internationalization. Section 3 develops the specific framework of this study. Sections 4 and 5 discuss, respectively, the data and the methodology. Section 6 presents the results and the last section summarizes and concludes.
2. Theoretical background
Innovation is the result of many factors operating at the macro, meso and micro levels. One element overarching all three levels of aggregation is internationalization. It has been claimed that companies that operate in many countries learn from different innovation contexts and are therefore able to benefit from them (e.g., Castellani and Zanfei 2006; Dunning and Wymbs 1999; Frenz and Ietto-Gillies 2007; Simard and West 2006). The sources of learning and knowledge acquisition can be many. If a country is highly internationalized it is likely to have a higher innovation performance (Amendola et al. 1993; Carlsson 2006; Filippetti et al. 2009) because: (i) its resources, its products and its institutions are exposed to alternative innovation contexts, and this allows firms and people to learn from different environments; and (ii) competition forces the firms to innovate.
Knowledge transmission mechanisms can be many and involve relationships between customers and sellers, principal and contractors, academic research networks, or employees working for the various institutions or moving between different employers (Frenz and Ietto-Gillies 2009; Laursen and Salter, 2004; Tether 2002). These mechanisms operate at both the national and international levels. At the international level transnational companies (TNCs) have a specific and additional transmission mechanism that operates via the internal network of the company. Knowledge is transmitted through the contacts between each unit of the TNC (be it subsidiary or headquarter) and also exchanged with the local environments in which the units operate.
The link between innovation and internationalization has a strong theoretical underpinning. The evolutionary theory of the firm ADDIN EN.CITE Nelson198241416R. Nelson S. Winter An Evolutionary Theory of Economic Change1982Cambridge, MAHarvard University PressNelson199342425R. NelsonN. Rosenberg R NelsonTechnical Innovation and National SystemsNational Innovation Systems: A Comparative Analysis1993OxfordOxford University Press(Nelson and Winter 1982; Nelson and Rosenberg 1993) has led to new developments in the theory of TNCs ADDIN EN.CITE Cantwell19896656656Cantwell, JTechnological Innovation and Multinational Corporations1989Oxford BlackwellKogut199366666617Kogut, BZander, U Knowledge of the firm and the evolutionary theory of the multinational corporationJournal of International Business Studies Journal of International Business Studies6256452441993(Cantwell 1989; Kogut and Zander 1993) in which the behaviour and activities of TNCs are linked to innovation development and diffusion.
TNCs operate in foreign countries through several modalities ranging from foreign direct investment (FDI) to trade to licensing to franchising to sub-contracting and to joint ventures. All modalities, in different ways, give rise to a variety of networks across countries. All these networks create scope for the acquisition of knowledge and innovation from diverse environments. The mechanisms through which the diffusion takes place can be via movements of tangible or intangible products, materials and assets or via the mobility of human resources.
Many companies particularly the large ones are organized in units operating in different localities. The company group is organized as an internal network and in the case of TNCs the units are located in different countries. Each unit of the TNC has the opportunity to learn from the innovation context and system in the country in which it operates be it the home or a host country. The knowledge is absorbed by the unit and then transmitted wholly or partially to other parts of the company via its internal networks ADDIN EN.CITE Zahra200066766717Zahra, S AIreland, R DHitt, M AInternational expansion by new venture firms: international diversity, mode of market entry, technological learning and performanceAcademy of Management JournalAcademy of Management Journal9259504352000Castellani200466866817Castellani, DZanfei, AChoosing international linkage strategies in the electronic industry: the role of multinational experienceJournal of Economic Behavior and OrganizationJournal of Economic Behavior and Organization4474755342004Castellani20066696696Castellani, DZanfei, AMultinational Firms, Innovation and Productivity2006CheltenhamEdward ElgarFrenz200767767717Frenz, MIetto-Gillies, GDoes multinationality affect the propensity to innovate? An analysis of the third UK Community Innovation Survey International Review of Applied EconomicsInternational Review of Applied Economics991172112007Frenz200567067017Frenz, MGirardone, CIetto-Gillies, GMultinationality matters in innovation. The case of the UK financial servicesIndustry and InnovationIndustry and Innovation65-921212005(Castellani and Zanfei 2004, 2006; Frenz et al. 2005; Frenz and Ietto-Gillies 2007, 2009; Zahra et al. 2000). Moreover, each company unit develops linkages with local businesses on innovation-related activities leading to external networks some of which are contractually formalized and others are more informal. There is, therefore, a double network contributing to knowledge and innovation acquisition and diffusion and thus to the capabilities of a specific country ADDIN EN.CITE Hedlund19906716715Hedlund, GRolander, DBartlett, C ADoz, YHedlund, GAction in heterarchies: mew approaches to managing the MNCManaging the Global Firm1990London RoutledgeCastellani200466866817Castellani, DZanfei, AChoosing international linkage strategies in the electronic industry: the role of multinational experienceJournal of Economic Behavior and OrganizationJournal of Economic Behavior and Organization4474755342004Castellani20066696696Castellani, DZanfei, AMultinational Firms, Innovation and Productivity2006CheltenhamEdward Elgar(Castellani and Zanfei 2004, 2006; Hedlund and Rolander 1990): the network within the TNC, and the external network(s) between each unit of the TNC and the local businesses with which it interacts. Each unit of the TNC contributes to spillover effects in the country of operation through knowledge and innovation developed by itself and through what it has acquired via the internal network of the TNC of which it is part. The extent of knowledge diffusion via internal and external networks may partly depend on the internal organizational structure of the company ADDIN EN.CITE Hedlund198667267217Hedlund, GThe hypermodern MNC a heterarchy?Human Resource Management Human Resource Management9352511986Bartlett1989986736Bartlett, C AGhoshal, SManaging Across Borders: The Transnational Solution1989Boston, MAHarvard Business School PressGupta199167467417Gupta, A KGovindarajan, VKnowledge flows and the structure of control within multinational corporationsAcademy of Management Review Academy of Management Review7687921641991Hedlund198667267217Hedlund, GThe hypermodern MNC a heterarchy?Human Resource Management Human Resource Management9352511986Gupta200067567517Gupta, A KGovindarajan, VKnowledge flows within multinational corporationsStrategic Management JournalStrategic Management Journal4734962142000(Bartlett and Ghoshal 1989; Gupta and Govindarajan 1991, 2000; Hedlund 1986). However, the direct activities of TNCs may be only one way in which companies, institutions, people and countries come in contact with the innovation context of other economies.
Over and above the acquisition of innovation capabilities via the operations of TNCs, there is also acquisition via the operations of other actors. Trade most, though not all of which, originates with TNCs contributes to the acquisition of innovation capabilities by exposing domestic businesses to the needs of foreign clients or to their new products and processes. Imports increase a countrys innovation potential when relevant technological knowledge is embedded in machinery and equipment. In addition to and combination with embedded knowledge and codified knowledge, tacit knowledge plays a crucial role in bringing about innovation (Polanyi 1966). In particular with respect to the latter the international movements of highly skilled labour (Salt 1991, 1997; OECD, 2002) some internally to TNCs are key mechanisms in knowledge and innovation transfers. Moreover, cross-border collaborations between companies, academic institutions and individual researchers contribute to innovation capabilities, so do international academic exchanges and trainings. These various elements separately and in combination point to the existence of a causal relationship between internationalization and innovation not only at the level of TNCs but also at the level of countries.
Most of the above discussion assumes that internationalization affects (positively) innovation. However, the causal link could go the other way round: firms and countries that are innovative are more likely to be able to conquer international markets and/or take up investment opportunities in foreign locations. The literature on the causal links between technological innovation and internationalization goes back many decades. Posner (1961) and Hufbauer (1966) linked trade performance to the technology gap between countries. Vernon (1966) extended the linkage to the impact of innovation on international production.
In reality there is likely to be a two-way interactive process in which innovation and internationalization reinforce each other leading to cumulative effects. A virtuous (or vicious) circle sets in. Innovative firms are more successful in international business. This puts them into contact with alternative business cultures and innovation contexts, thus adding to their overall business knowledge. This in turn makes them more innovative and thus more able to compete internationally. Less innovative firms and countries may become locked into an opposite vicious circle.
This research is about association not causality. The above arguments and literature lead us to support the view that significant correlations between measures of innovation and internationalization are not spurious but are a sign of underlying causality. However, correlation coefficients in themselves do not give us guidance as to the possible direction of causality particularly because there are theoretical arguments for causation to go either way. We therefore keep an open mind as to the direction of the relationship between innovation and internationalization. The reader should bear in mind that, whatever the main direction of causation, a virtuous or vicious circle in innovation and internationalization is likely to be at work for innovative and non-innovative countries respectively.
3. The framework
Internationalization can take place via different modalities, some of which are determined by the activities of TNCs as discussed above. The modalities of internationalization considered here span wider than the activities of TNCs. They are: inward and outward FDI; trade (both imports and exports); cross-border influx of skilled personnel and of students. The inclusion of both inflows and outflows for FDI and trade respond to the assumption that firms learn from their contacts with other business units in foreign countries in any type of business contacts, be they as buyer or seller, recipient or initiator of cross-border investment and trade.
We test whether the association between internationalization and innovation holds for different levels with respect to the industry sectors considered: for the level that includes all industries (A); and for the level of technology-intensive industries only (B). In other words, does it matter for the association between internationalization and innovation that a country may be ranking low on the overall aggregate internationalization if it comes high on the internationalization ranking for innovation-intensive industries and vice versa?
For levels A and B we use national statistics. We also use data from two surveys Innobarometer and CIS to conduct further analyses at what we call level C. These two surveys allow us to calculate indicators of countries internationalization expressed as the share of enterprises that reported cross-border activities in each country. The aims of the level C analysis are twofold: first, to consider additional dimensions of internationalization made possible by the use of variables available from the surveys but not from official statistics (e.g. cross-border collaborations); second, to check the robustness of our results at levels A and B. We can therefore examine whether there is consistency between results based on the international embeddedness of all industries and/or for high-tech industries only, and those that emerge from the firms themselves.
The internationalization framework we shall be working with is one that stresses the overall level of activities abroad or from abroad. Other dimensions of internationalization such as the spatial dimension will not be taken into account. For example the proximity of countries or the number of countries with which any one country has economic relationships will not be considered. In other words we concentrate on what has been labelled as the intensity dimension of internationalization rather than its extensity/spatial dimension, though the latter may also be important for the relationship between internationalization and innovation (Ietto-Gillies 2009).
The plan for the research is therefore to assess the innovation performance of countries as measured by the SII against three sets of variables capturing the degree of internationalization. The degree of internationalization is here interpreted to mean the degree to which the country is open to operations with all foreign countries considered together, independently of distance spatial and/or cultural or number of foreign countries.
The first set of country-level variables is derived from general indicators of internationalization (e.g. FDI flows for all industries together); the second set relates to indicators based on data for innovation-specific industries; the third set refers to country-level indicators based on the share of enterprises that reported international activities in Innobarometer and CIS. Box 1 details the indicators and the composite indices for our three sets.
Box 1 here
4. Data
In this section we discuss the data chosen for the innovation indicators and then those collected for the internationalization indicators. The methodology is discussed in the next section. The countries considered in our analysis are those for which a specific set of data on innovation already exists: the European Innovation Scoreboard Summary Innovation Index (SII). Specifically, they are the EU27 countries plus Croatia, Turkey, Iceland, Norway and Switzerland. With respect to the timeframe, innovation scores relate to the years 2004 to 2008, while internationalization is measured from 1999 to 2007, with the exception of level C which is based on one cross-section of the relevant surveys. The Innobarometer data refers to the period 2006-2008 and the CIS data to the period 2002-2004.
4.1 Innovation data
The SII is compiled and published by European Commission since 2001. Changes in the methodology applied to different waves of the EIS mean that comparable data on SII is available only from 2004 onwards ADDIN EN.CITE European Commission200959259227European Commission,European Innovation Scoreboard 2008. Comparative analysis of innovation performance2009BrusselsEuropean Commission, DG Enterprise(European Commission, 2009a). The SII is an aggregate index and is based on 29 individual variables as reported in Appendix I. The SII captures innovation performances of countries and is based on measures such as the share of innovators in a country or the average turnover from innovations, but the EIS also covers wider framework conditions, such as finance and support for innovation, human resources and ICT infrastructures. The 29 variables feeding into the SII are grouped in the Innovation Scoreboard into seven sub-categories: human resources, finance and support, firm investment, linkages and entrepreneurship, throughputs, innovators and economic effects.
We test for associations of measures of internationalization and the SII. Additionally, we test for associations with two subsets of SII. These subsets are based on variables that are direct inputs into innovation or direct outputs. Out of the 29 variables we selected as follows. Input variables: business R&D expenditures as a percentage of GDP, IT expenditures as a percentage of GDP, and (non-R&D) innovation related expenditures by firms as a percentage of GDP. Output variables: average new-to-market sales as a percentage of total sales, new-to-firm sales as a percentage of total sales, share of firms using patents, trademarks, or registered designs.
4.2 Internationalization data
To assess the extent to which the internationalization of countries and their innovation performance are associated we first identify three levels at which we want to consider internationalization: level A, B and C as discussed in section 1. For each level several variables are considered as listed in Box 2.
Box 2 here
Specifically, level A includes the following variables: inward and outward FDI, imports and exports, mobility of employees and of students. Level B includes inward and outward FDI for innovation-intensive manufacturing sectors and for knowledge-intensive services, imports and exports of innovation-intensive products, balance of payments debits and credits for knowledge-intensive services, and mobility of research students.
Under level C we include those questionnaire items in the Innobarometer and CIS surveys that have a bearing on the international embeddedness or focus of responding companies. Specifically, the following variables are taken from Innobarometer (CI): proportion of enterprises that operated in international markets, outsourced activities to firms located abroad, invested in firms located abroad, cooperated with partners which were located abroad, recruited employees from other countries, carried out market-testing in foreign countries, considered international markets to be the lead markets. The following are derived from the Community Innovation Survey (CII): proportion of enterprises that operated in international markets, proportion of foreign-owned enterprises and proportion of enterprises reporting cooperation with partners abroad.
For levels A and B, a variety of official data sources are used, including the United Nations Conference on Trade and Development for FDI data, the World Development Indicators produced by the World Bank for data on trade, the EU Labour Force Survey for the number of total and foreign employees and the Education Statistics from the Organisation of Economic Cooperation and Development for students and GDP. We collected the raw data from these different sources for the years 1999 up to 2007 the latest available year.
With respect to level C, the data refers to one reference period: activities between 2006 and 2008 in the case of Innobarometer, and between 2002 and 2004 in the case of CIS4. In fact, the relevant internationalization questions of the Innobarometer survey are only available for this latest wave of Innobarometer. The number of countries included at level C is less than those for levels A and B. It is, respectively, 29 for the analysis from Innobarometer data (level CI) and 28 for the one using CIS data (level CII). The following countries included in levels A and B are missing: for level CI Croatia, Iceland and Turkey; for level CII Croatia, Switzerland, Turkey and the UK.
5. Methodology
In this section we first discuss how the indicators and summary indices for levels A and B data are developed. This is followed by the discussion on level C. When computing the individual indicators for levels A and B, we smooth the data by using five year moving averages. We cumulate both the nominators and denominators of the internationalization indicators over five years. For example, we sum the values for FDI inflows for 1999 to 2003 and express the total over the five years as a percentage of GDP cumulated over the same period. Thus, between 1999 and 2007, there are five consecutive indicators, the first referring to the period 1999 to 2003 (our first time period) and the last to the period 2003 to 2007. This smoothing is done for two reasons. Firstly, because flow data such as the data on FDI is subject to some degree of volatility and this is flattened through the use of moving averages. Secondly, to capture in the indicators a cumulative process of learning by which a countrys innovation performance is not only affected by the level of international embeddedness in the same or previous years, but depends on the cumulative impact of international linkages and learning over a period of time.
The six variables in A1 to A3 (Box 2 section A) are expressed in relative terms, i.e. as a percentage of GDP, total number of employees or total number of students. The variables differ considerably in terms of their average values; for example trade expressed as a proportion of GDP typically takes values in the region of 0.5, while FDI expressed as a proportion of GDP takes values of around 0.05. Moreover, the variables differ across countries according to the size of the country and the structure of its economy. We normalized the variables and turn them into indicators that range from 0 to 1, partly to offset the problems of scale just mentioned but also to provide a reliable comparison between our indices and the SII we use to capture innovation. The latter uses the same method to compute indicators of SII. The normalization was done as follows:
EMBED Equation.3 (1)
with i denoting the 32 countries and t the five time periods.
While computing the indicators, we inspected and adjusted the raw data (variables) for outliers using the interquartile range (IQR). IQR is equal to the distance between the first and third quartiles (or between the 75th and 25th percentiles): IQR=Q3-Q1, where this distance spans the middle 50% of the data. Outliers are identified as follows: negative outliers are values below Q1 3IQR, while positive outliers are greater than Q3 + 3IQR. Outliers are not included in determining the maximum and minimum scores in the normalization process. For outliers where the value of the relative score is above the maximum score or below the minimum score the re-scaled score is set to 1 and 0 respectively.
From the six indicators (related to the six variables included in A1-A3), we compute a summary index SGI/A as the average of the six indicators. For some countries not all of the six variables and thus indicators are available. In such instances the average is calculated over the available indicators. A similar procedure is followed for level B for which we consider nine variables (see Box 2 section B).
The SGI/A and SGI/B and their underlying indicators of internationalization are linked with a time lag of one year to the SII; the SGI/A and B based on 1999 to 2003 data are associated with the SII for the year 2004, the SGIs based on 2000 to 2004 data to SII for 2005 and so forth up to SII for 2008. The rationale for the lag is that international embeddedness in the earlier time period feeds into the innovation performance of a country in the later period. In the case of level C the SGI/C is available for the reference period of the surveys (2006 onwards in the case of the Innobarometer and 2002-2004 in the case of CIS) and the indicators are correlated with the latest SII (2008).
Turning now to level C, the Innobarometer survey used in this paper was conducted in April 2009. It is the first wave of the survey that contains a range of questions specifically aimed at measuring internationalization. Questionnaires were completed by enterprises in 29 European countries, with a total of 5,234 responses. The unit of analysis is the enterprise, which is the smallest independent reporting unit which may be part of a much larger company group, and is often located in a single site, but can comprise more than one site. Each of the 29 countries achieved a sample size of 200 with the exceptions of Norway and Switzerland for which there are 100 observations, and Cyprus, Luxembourg and Malta with 70 observations. The survey contains seven questionnaire items that relate to aspects of international embeddedness of firms activities listed in Box 2 under CI. All items provide binary data indicating whether or not enterprises engaged in the relevant international activities.
We compute the proportion of firms that, for example, operated in international markets (compared with all firms that responded to the survey). We then follow the same data transformation to derive at seven indicators ranging from 0 to 1 (see Equation 1 above). Finally, we calculate a level CI globalisation index that is the simple average of the seven indicators. The reference period for the Innobarometer indicators is 2006 to 2008, and this is linked to EIS data for 2006 and 2007.
Additionally to the Innobarometer indicators, we use the fourth European CIS, conducted by the individual member states and compiled by Eurostat for 27 countries (while Switzerland and the UK conduct CISs, they do not deposit the data in the Eurostat database). As with the Innobarometer, the unit of analysis is the enterprise. The reference period is 2002-2004. While two of the variables used enterprise operated in international markets and enterprise cooperated internationally are also captured by the Innobarometer survey, the other variable foreign-ownership is not. The number of observations that feed into the EU CIS is much larger just below 70,000 enterprises compared with the Innobarometer, so the results act as a robustness check for the Innobarometer results.
At each level of analysis partial correlation coefficients between SII, SII-inputs and SII-output, and the internationalization indicators are calculated for the single indicators that compose each set at the three levels (A, B and C) and the aggregate Summary Globalization Index (SGI) for each specific level. The data at levels A and B are pooled for all available periods.
The correlation coefficients are partial correlations controlling for: (i) the natural log of the population of countries; (ii) the gross fixed capital formation cumulated for five years over GDP as a proxy for capital intensity; and (iii) the surface area of each country measured in squared kilometres. These partial correlations do not differ substantially from the zero-order correlations which we computed but do not present. Additionally, and in order to increase the robustness of our findings, we calculated partial correlations using alternative measures of the size of countries, including GDP per capita. Again this did not substantially alter the results.
Over and above the use of control variables, we grouped countries into small versus large countries and computed correlations on these two subgroups. Cluster analysis is used to group countries. Again this did not impact on the overall findings. Filippetti et al. (2009) gives details of the two groupings and the related results. It also gives the computed indicators and indices for all the countries.
6. Results
This section provides the results of the correlation analyses organized into three sub-sections addressing levels A, B and C respectively.
6.1 Level A correlations between internationalization and innovation
Table 1 provides the partial correlations of SGI/A and the six internationalization indicators of level A, with the overall innovation index SII, and with the two narrower innovation indices: SII-inputs and SII-outputs. The latter two indices are calculated for selected EIS variables as indicated in the last paragraph of section 4.1. All results presented are based on the pooled dataset.
Table 1 here
The positive correlations between SII, SII-inputs and SII-outputs with SGI/A, outward FDI flows, exports, foreign students and foreign employees are all statistically significant. The strong associations between innovation and outward FDI points to two possible explanations: (i) innovative firms and countries are the ones that compete successfully in taking up investment opportunities abroad; or (ii) linkages via foreign investment and foreign subsidiaries allow TNCs to learn from other business cultures and this affects positively the innovation score of their home country. The causality could go in either direction.
The strong results for foreign students and foreign employees point to the following. Skilled foreign human resources bring knowledge, specifically tacit knowledge, into the host country with positive effects on innovation. Conversely, innovative countries attract skilled foreign human resources.
The positive correlations between exports and innovation, in particular with reference to SII-inputs, point towards the following relationship; countries whose firms spend more on innovation-related activities are able to compete in international markets and therefore export more.
Looking at the first row of Table 1 we see that the results show a higher correlation coefficients between SII-inputs and SGI/A (r=0.59; p<0.01) and between SII-outputs and SGI/A (r=0.52; p<0.01) compared with what we obtained for the correlation between SII and SGI/A (r=0.46; p<0.01). With respect to the six individual internationalization indicators, all six of them, with the exception of inward FDI flows, are positively and significantly associated with SII-inputs, while outward FDI flows, exports, foreign students and foreign employees are significantly associated with SII-outputs.
We also computed the correlations between SGI/A and SII without the time-lag and the smoothing of data, thus correlating SII2004 with SGI/A based on 2004 data and so on. Fewer data pairs are available for these correlations. Matching years for which we have internationalization and innovation data are 2004, 2005, 2006 and 2007. The results of the correlations based on 2004 to 2007 data without time-lag are highly similar. The pooled correlation coefficient between SII and SGI/A is r=0.33 (p<0.01) compared with r=0.46 (p<0.01) from the Table 1.
To further explore the possibility of a two-way causation between innovation and internationalization, we computed the correlations presented in Table 1 with reverse time-lags. The available data pairs relate to the following years: innovation in 2004 and internationalization in 2005, innovation in 2005 and internationalization in 2006, and finally innovation in 2006 and internationalization 2007. As is the case for the correlations without time-lag, the data on internationalization is not smoothened over a five-year period as we do not have enough data points to do so. Again, we find highly similar patterns: both in terms of the pairs of indicators that remain significantly associated (SGI/A, outward FDI, foreign students, foreign employees, exports); and the strengths of the association (i.e. the size of the correlation coefficients). Therefore, the pattern is in line with the virtuous/vicious circle between innovation and internationalization proposed in section 2.
6.2 Level B correlations between internationalization and innovation
Table 2 provides the correlations between SGI/B and 9 separate indicators of internationalization with SII, SII-inputs and SII-outputs. The results for level B reinforce those of level A: innovation and internationalization appear to be correlated at both overall level and at the level of innovation-intensive industries, and specifically with respect to outward FDI, movement of foreign students and exports.
Table 2 here
The correlation coefficients computed at level B are not higher compared with the level A analysis. This seems to indicate that internationalization per se and not necessarily with respect to high technology and knowledge intensive industries/products impacts on countries innovation performance. We might have expected to see stronger associations with the innovation scoreboard at level B. However, it is worth noting that correlations do not describe size effects but are measures of co-variation that express the closeness of pairs of scores around either a positive or negative straight line. The slope of the straight line the size effect is not reflected in the correlation coefficient, but would require different estimations.
Looking at the results for both level A and B together, there are two sets of variables we have not yet commented on: inward FDI and imports. In these cases the results are more uncertain compared with the results for outward FDI, for exports and for the variables related to the mobility of human resources.
With respect to inward FDI the relationship between this variable and innovation is not straightforward. The existence of inward FDI into a country may be a sign that foreign TNCs are more competitive than domestic firms, partly by being more innovative than the domestic firms. This is, indeed, how we interpret the strong results we get for correlations with outward FDI. However, there is also some evidence that, in the case of specific countries such as the UK, foreign TNCs may be attracted to invest in countries with a strong innovation environment (Driffield and Love, 2003). There is also evidence that innovation and technological spillovers from inward FDI into the host country may be connected to a variety of elements ranging from the nature of FDI (horizontal or vertical) to R&D intensity (Castellani and Zanfei, 2006: Part III). The uncertainty of the results for inward FDI may be the outcome of these conflicting forces at work. The true situation may require country by country studies.
With respect to imports, we find low but significant correlations with SII, SII-inputs and SII-outputs at level B. There may be an indication here that countries with a strong innovation base are also the ones that have larger volumes of imports, including imports of high technology products, and, in turn, imports in high-tech products are likely to impact positively on the innovation performance and thus on the export performance with respect to high technology products as well as on the need to import further high-tech products in a cumulative virtuous process.
Another possible interpretation of these correlations can be found via the link with absorptive capacity. Cohen and Levinthal (1989, 1990) convincingly argue that investment in R&D not only generates information but it also increases the firms ability to absorb knowledge that may spill over from other firms. We are here suggesting the following possible mechanisms with respect to imports of high-tech products. Absorptive capacity is higher in countries that import such products because: (a) the firms involved in such imports enhance their internal capabilities in order to successfully integrate and use products involving new technologies; or/and (b) imported high-tech products themselves enhance the absorptive capacity. Either case or a combination of both would lead to the positive associations between imports of high-tech products and SII-inputs (r=0.23; p<0.05) as well as SII-outputs and exports (r=0.40: p<0.01).
The three variables that feed into SII-inputs (business R&D expenditure, IT expenditure, and non-R&D innovation expenditure) are all indicators of formation of innovative capabilities and this in itself explains why both imports and exports are associated with SII-inputs more definitely (at both levels A and B) than SII-outputs or SII. We correlated the individual variables business R&D, IT expenditure and non-R&D innovation expenditure with imports and exports of high-tech products. IT expenditure is the variable with the strongest association with imports (r=0.21; p<0.05) and exports (r=0.23; p<0.01). While business R&D is associated only with exports (r=0.30; p<0.01), thus pointing more strongly towards explanation (b) above. Non-R&D innovation expenditure is uncorrelated with both imports and exports.
6.3 Level C correlations between internationalization and innovation
Finally, in Table 3 we present the results for level C. Indicators are based on the international activities as reported by enterprises responding to the Innobarometer and CIS4 surveys. We analyse separately indicators related to these two surveys under CI and CII respectively.
Table 3 here
The summary index SGI/CI and SII have a correlation of 0.45 (p<0.05), and SGI/CI and SII-inputs a correlation of 0.37 (p<0.1), similar to the correlation coefficients for SGI/A and SGI/B; however, SGI/CI is not associated with SII-inputs. The strongest associations with the innovation scores among the individual indicators at level C are found with the share of enterprises recruiting employees from other countries, investing abroad and market testing abroad. The first two variables are similar to the level A indicators on share of foreign employees and outward FDI.
The correlations between the indicators derived from CIS4 and SII (presented under CII in Table 3) are somewhat larger which might be the case because some of the variables feeding into the SII index are derived from different sections of CIS and, therefore, are derived from the same enterprises. The strongest correlations link the share of foreign-owned enterprises in a country with SII, SII-inputs and SII-outputs. The CIS does not have information on whether the surveyed enterprise is part of a domestic TNC with investment abroad. Thus a variable related to outward FDI consistent with that considered under A, B and CI is not available within CII.
Nonetheless, looking at the results across CI and CII (specifically those for indicators: investment into firms located abroad from CI and enterprise is foreign-owned from CII) we can observe the following. Countries whose enterprises are part of a TNC be it domestically or foreign-owned appear to have stronger association with innovation. This result is consistent with other studies in which multinationality was found to affect innovation positively independently of whether the TNC is a foreign or domestic company (Frenz and Ietto-Gillies 2007, 2009; Castellani and Zanfei 2004) for the reasons explained in section 2.
Fairly significant are also the results for enterprise operates in international markets but not for the CI and CII indicators enterprise is involved in cooperation with partners abroad. The latter result is consistent with those obtained at the firm level in Frenz and Ietto-Gillies (2009). It may be due to issues of appropriability of knowledge relevant for innovation, and to the exchange of knowledge, specifically tacit elements of knowledge, across borders but outside the company environment.
7. Summary, conclusions and future research
The paper tests the association between internationalization and innovation for the 32 countries for which the EIS innovation scores are available. It starts by recalling some of the theoretical background to as well as empirical evidence for the relationship between internationalization and innovation. It then presents the framework underlying the assessment of the relationship between innovation and internationalization. The empirical work all addressed at the country level explores correlations between innovation and several indicators of internationalization at three levels of analysis. The full aggregate level (A) in which internationalization variables are considered for the whole country and all industries; at level B of the analysis, the study focuses on how the share of such industries within a country affects the strength of its link between innovation and internationalization.; and (C) the share of firms reporting international activities in each country based on data from two surveys the Innobarometer and CIS.
Russo and Williamson (2007) convincingly argue that establishing causation in the health sciences involves both of the following: (a) evidence of probabilistic/statistical association between relevant variables; and (b) evidence of mechanisms through which variable(s) X affect(s) variable Y. In this paper we (a) provide evidence of statistical association and (b) put forward possible mechanisms on the basis of theoretical arguments and evidence form other studies. The combination of the two together is, in our view, strong evidence that the correlations are not spurious but indicative of causality.
However, it is not possible from our evidence to draw inference on the direction of causation. Causality could go from innovation to internationalization or vice versa. If a country is highly internationalized it is likely to have a high innovation performance. Conversely, firms and countries that are innovative are more likely to be able to penetrate international markets and/or take up investment opportunities in foreign locations. In reality a virtuous (or vicious) circle is likely to set in. Innovative firms are more successful in international business. This puts them into contact with alternative business cultures and innovation contexts, thus adding to their overall business knowledge. This in turn makes them more innovative and thus more able to compete internationally. Less innovative firms and countries may become locked into an opposite vicious circle.
The international variables that show clear association throughout are those related to outward foreign direct investment and similar indicators of firm behaviour, foreign students and foreign employees. The strong associations obtained for the latter two variables indicate the relevance of cross-country mobility of skilled human resources for the acquisition of knowledge and innovation capabilities.
The results for imports, exports and for inward FDI are more uncertain. Regarding trade the results are, however, all positive and significant for level B internationalization (focusing on innovation-intensive industries). Innovative countries with a relative specialization in high-tech sectors import high-tech products; this supports their innovation performance and helps them to compete abroad via the exportation of high-tech products (which are not necessarily part of the same industrial category as the imports).
There are clear limitations in a study of this nature. In terms of data they range from: the composition of the EIS itself to the timescale set by the availability of the EIS as well as of the two surveys (CIS and Innobarometer). At the theoretical level there are also limitations such as: the inability to capture the direction of causality; and the inability to take into account further dimensions of internationalization and specifically the spatial/extensity dimension (as highlighted in section 3). In terms of causality there is also the inability to take account of other variables affecting innovation or internationalization.
Nonetheless even within the boundaries of this exploratory research some definite conclusions are possible based on the empirical results as well as on the underlying theoretical framework. Firstly, the persistence of significant results for outward FDI and for the variables related to mobility of human resources points to the relevance of internationalization for innovation. Secondly, the overall results point to interactive effects between innovation and internationalization with possible cumulative virtuous or vicious effects.
Overall the results are robust and interesting enough to warrant further work in the following areas. First, the building up of evidence on the direction of causality. Second, the assessment of the strength of causal relationships in models that take account of other explanatory variables.
References
Amendola, G., Dosi, G., Papagni, E. 1993. The dynamics of international competitiveness. Review of World Economics, 129 (3): 451-471.
Archibugi, D., Pianta, M. 1992. The Technological Specialization of Advanced Countries. Kluwer Academic Publishers, Dordrecht.
Bogliacino, F. and Pianta, M. 2009. Innovation performances in Europe: a long term perspective. Thematic paper for the European Commission Pro Inno Europe.
Bartlett, C. A., Ghoshal, S. 1989. Managing Across Borders: The Transnational Solution, Harvard Business School Press, Boston, MA.
Cantwell, J. A. 1989. Technological Innovation and Multinational Corporations, Blackwell, Oxford.
Cantwell, J. A. 1994, The relationship between international trade and international production, in D. greenway and L. A. Winters (eds), Surveys in International Trade, Oxford: Basil Blackwell.
Cantwell, J. and F. Sanna Randaccio, 1993. Multinationality and firm growth, Weltwirtschaftliches Archiv, 129 (2): 275-99.
Carlsson, B. 2006. Internationalization of innovation systems: A survey of the literature. Research Policy 35 (1): 56-67.
Castellani, D., Zanfei, A. 2004. Choosing international linkage strategies in the electronic industry: the role of multinational experience. Journal of Economic Behavior and Organization, 53 (4), 447475.
Castellani, D., Zanfei, A. 2006. Multinational Firms, Innovation and Productivity. Edward Elgar, Cheltenham.
Cohen, W. M. and Levinthal, D. A. 1989. Innovation and Learning: The Two Faces of R&D. Implications for the analysis of R&D Investment. Economic Journal, September, 569-596.
Cohen, W. M. and Levinthal D. A. 1990. Absorptive Capacity: A New Perspective on Learning and Innovation Administrative Science Quarterly, 35, 128-195.
Driffield, N., Love, J. H. 2003. FDI, technology sourcing and reverse spillovers. The Manchester School, 71 (6): 659-672.
Dunning, J.H. and Wymbs, C. 1999. The geographical sourcing of technology-base assets by multinational enterprises, in Archibugi, D., Howells, J. and Michie, J. (eds) Innovation Policy in a Global Economy, Cambridge: Cambridge University Press, 184-224.
European Commission 2009a. European Innovation Scoreboard 2008. Comparative analysis of innovation performance. Brussels, European Commission, DG Enterprise and Industry.
European Commission 2009b. Innobarometer 2009. Brussels, DG Enterprise and Industry.
Fagerberg, J. 1996. Technology and competitiveness. Oxford Review of Economic Policy, 12 (3): 39-51.
Filippetti, A, Frenz, M and Ietto-Gillies, G. 2009. Is the innovation performance of countries related to their internationalization. Thematic paper for the European Commission Pro Inno Europe.
Frenz, M., Girardone, C., Ietto-Gillies, G., 2005. Multinationality matters in innovation. The case of the UK financial services, Industry and Innovation 12 (1): 65-92.
Frenz, M., Ietto-Gillies, G. 2007. Does multinationality affect the propensity to innovate? An analysis of the third UK Community Innovation Survey, International Review of Applied Economics, 21 (1): 99117.
Frenz, M., Ietto-Gillies, G. 2009. The impact on innovation performance of different sources of knowledge: Evidence from the UK Community Innovation Survey. Research Policy, 38 (7): 1,125-1,135.
Grimwade, N. 2000. International Trade: New Patterns of Trade, Production and Investment. London: Routledge, 2nd edition.
Grubel, H.G. and Lloyd, P. J. 1975. Intra-Industry Trade. London: Macmillan.
Gupta, A. K., Govindarajan, V. 1991. Knowledge flows and the structure of control within multinational corporations, Academy of Management Review, 16 (4): 768792.
Gupta, A. K., Govindarajan, V. 2000. Knowledge flows within multinational corporations, Strategic Management Journal, 21 (4): 473496.
Hedlund, G. 1986. The hypermodern MNC a heterarchy? Human Resource Management, 25 (1): 935.
Hedlund, G., Rolander, D. 1990. Action in heterarchies: new approaches to managing the MNC, in Managing the Global Firm by C. A. Bartlett, Y. Doz, G. Hedlund (eds), Routledge; London.
Hufbauer, G. C. 1966. Synthetic Materials and the Theory of International
Trade. London: Duckworth.
Hughes, K 1986. Exports and innovation: A simultaneous model, European Economic Review 30 (2): 383-399.
Ietto-Gillies, G. 2009. Conceptual issues behind the assessment of the degree of internationalization, Transnational Corporations, 18 (3): 59-83.
International Monetary Fund 1993. Balance of Payments Manual. Fifth edition.
Kafourosa M.I., Buckleya P.J., Sharpb J. A. and Wangc, C. 2008. The role of internationalization in explaining innovation performance. Technovation, 28 (1-2): 63-74.
Kleinknecht, A., Oostendorp, R. 2002. R&D and export performance: taking account of simultaneity, in A. Kleinknecht, Mohnen P. (eds.), Innovation and firm performance, London: Palgrave, pp. 310-320.
Kogut, B., Zander, U. 1993. Knowledge of the firm and the evolutionary theory of the multinational corporation. Journal of International Business Studies 24 (4): 625645.
Krugman, P. 1995. Technological change in international trade. In Stoneman P. (ed.), Handbook of Economics of Innovation and Technological Change, Oxford, Blackwell.
Laursen, K., Salter, A. 2004. Searching high and low: what types of firms use universities as a source of innovation?. Research Policy 33 (8): 1,2071,215.
Nelson, R., Rosenberg, N. 1993. Technical Innovation and National Systems. National Innovation Systems: A Comparative Analysis. R. Nelson. Oxford University Press, Oxford.
Nelson, R., Winter, S. 1982. An Evolutionary Theory of Economic Change. Harvard University Press, Cambridge, MA.
Organization for Economic Cooperation and Development 2002. International Mobility of the Highly Skilled. Paris: OECD Publishing.
Polanyi, M. 1966. The Tacit Dimension. Doubleday, New York.
Posner, M.V. 1961. International trade and technical change. Oxford Economic Papers, 13: 323-341.
Russo, F. and Williamson J. 2007. Interpreting causality in the health sciences, International Studies in the Philosophy of Science, 21 (2): 157-170.
Salt, J. 1991. Migration processes among highly skilled in Europe. International Migration Review, 26: 484-505.
Salt, J. 1997. International movement of the highly skilled, Directorate for Education, Employment, Labour and Social Affairs. International Migration Unit: Occasional Paper no. 3, OECD/GD (97), 169.
Simard, C. and West J. 2006, Knowledge networks and the geographic locus of innovation, in Chesbrough, H., Vanhaverbeke, W. and West, J (eds.) Open innovation. Researching a new paradigm. Oxford: Oxford University Press.
Tether, B.S. 2002. Who co-operates for innovation, and why: an empirical analysis. Research Policy, 31 (6): 947967.
Vernon, R. 1966. International investment and international trade in the product
cycle. Quarterly Journal of Economics, 80: 190-207.
Zahra, S. A., Ireland, R. D., Hitt, M. A. 2000. International expansion by new venture firms: international diversity, mode of market entry, technological learning and performance. Academy of Management Journal 43 (5): 925950.
Box 1 Overview of the different level of internationalization indicators and innovation indicator
Level A. Internationalization indicators: general levelFDITradeMobility of employees and students
Level B. Internationalization indicators: innovation-specific industry level FDI in hi-tech manufacturing and KIBS Trade in hi-tech manufacturing and KIBS Mobility of research students
Level C. Internationalization indicators: innovation survey data Relevant questionnaire items in Innobarometer and CIS including: international investment, international cooperation, outsourcing, foreign employees and foreign ownership
Box 2 Levels A to C internationalization indicators and data sources
Variables for levels A, B, CData source
A. Internationalization indicators: general level
A.1 FDI Inward FDI flows for all industries as % of GDPUnited Nations Conference on Trade and Development DatabaseOutward FDI for all industries as % of GDPA.2 Trade flowsImports as % of GDPWorld Development Indicator produced by the World BankExports as % of GDPA.3 Mobility of employees and studentsForeign students in tertiary education as % total students in tertiary educationEducation Statistics produced by the Organisation of Economic Cooperation and Development Foreign employees as % of total employeesEU Labour Force Survey collected by EurostatSummary Globalisation Index at level A (SGI/A)
B. Internationalization indicators: innovation-specific industry level
B.1 FDIInward FDI in high-tech manufacturing as % GDPOECD (International Direct Investment Statistics)Outward FDI in high-tech manufacturing as % GDPInward FDI in knowledge intensive services as % GDPOECD (Globalisation Statistics)Outward FDI in KIBS as % total FDIB.2 TradeImports of high-tech products as % of GDPEurostat (Science and Technology)Exports of high-tech products as % of GPDDebits in knowledge intensive services as % of GDPIMF Balance of PaymentsCredits in knowledge intensive services as % of GDPB.3 Mobility of research studentsForeign research students as % total research studentsOECD (Education Statistics)
Summary Globalisation Index at level B (SGI/B)
C. Internationalization indicators: innovation survey data
CICI.1 Enterprises operating in international markets as % of all companiesEuropean Commission (Innobarometer)CI.2 Enterprises that outsource to companies located abroad as % of all companiesCI.3 Enterprises investing abroad as % of all companiesCI.4 Enterprises that engaging in international cooperation as % of all companiesCI.5 Enterprises that recruit employees from abroad as % of all companiesCI.6 Enterprises that engage in market-testing abroad as % of all companiesCI.7 Enterprises that consider lead markets to be abroad as % of all companies Summary Globalisation Index at level CI (SGI/CI)CIICII.1 Enterprises that are foreign-owned as a % of all enterprisesEurostat (Community Innovation Survey)CII.2 Enterprises that operate in international markets as % of all enterprisesCII.3 Enterprises that cooperated with international partners as a % of all enterprisesSummary Globalisation at Index level CII (SGI/CII)
Table SEQ Table \* ARABIC 1 Level A results correlations between six internationalization indicators, SGI/A and SII, SII-inputs and SII-output, pooled data
Innovation indexSIISII-inputsSII-outputsInternationalization indicators and index (1) (2) (3)SGI/ACorrelation0.46***0.59***0.52***p-value0.000.000.00N153110150Inward FDI flowsCorrelation-0.030.140.04p-value0.740.150.62N153110153Outward FDI flowsCorrelation0.62***0.43***0.62***p-value0.000.000.00N153110150Imports Correlation-0.14*0.18*-0.05p-value0.090.070.5N151106146ExportsCorrelation0.14*0.37***0.19**p-value0.090.000.03N151106151Foreign studentsCorrelation0.67***0.72***0.65***p-value0.000.000.00N11387115Foreign employees Correlation0.54***0.52***0.60***p-value0.000.000.00N135104133All correlations control for the natural log of the population size of each country, for cumulative gross fixed capital formation as a percentage of GDP, and for the surface area in square km.
Significance level: * p<0.10; ** p<0.05; *** p<0.01
Table 2 Level B results. Correlations between nine internationalization indicators, SGI/B and SII, SII-inputs and SII-output, pooled data
Innovation indexSIISII-inputSII-outputInternationalization indicators and index (1) (2) (3)SGI/BCorrelation0.42***0.28***0.43***p-value0.000.000.00N150110150Inward FDI in high-tech manufacturingCorrelation-0.04-0.22*-0.11p-value0.700.080.28N936893Outward FDI high-tech manufacturingCorrelation0.43***0.45***0.26**p-value0.000.000.02N806080Inward FDI in knowledge intensive servicesCorrelation0.150.180.21*p-value0.180.180.06N866186Outward FDI in knowledge intensive servicesCorrelation0.150.020.20*p-value0.200.850.07N826182Imports of high-tech products
Correlation0.18*0.23**0.19**p-value0.030.020.02N150110150Exports of high-tech products
Correlation0.43***0.26***0.40***p-value0.000.010.00N150110150Debits in knowledge intensive servicesCorrelation0.26***0.17*0.27***p-value0.000.100.00N137100137Credits in knowledge intensive services Correlation0.31***0.22**0.31***p-value0.000.030.00N140100140Foreign research studentsCorrelation0.65***0.51***0.54***p-value0.000.000.00N916791All correlations control for the natural log of the population size of each country, for cumulative gross fixed capital formation as a percentage of GDP, and for the surface area in square km.
Significance level: * p<0.10; ** p<0.05; *** p<0.01
Table 3 Level C results. Correlations between seven internationalization indicators, SGI/C and SII2008
Innovation indexSIISII-input SII-output Internationalization indicators and index
Innobarometer (1)(2)(3)SGI/CI Correlation0.45**0.170.37*p-value0.030.470.07N272227Enterprise operated in international marketsCorrelation0.160.060.33p-value0.460.810.12N272227Outsourced activities to firms located abroadCorrelation0.07-0.13-0.01p-value0.750.590.94N272227Investment into firms located abroadCorrelation0.35*0.170.45**p-value0.090.490.03N272227Cooperated with partners which were located abroadCorrelation0.330.370.02p-value0.120.120.93N272227Recruited employees from other countriesCorrelation0.59**0.160.57***p-value0.020.500.00N272227Market-testing in foreign countriesCorrelation0.37*0.220.27p-value0.080.370.21N272227Enterprise considered international markets as lead marketsCorrelation0.09-0.130.03p-value0.680.610.88N272227CIS4SGI/CII Correlation0.45**0.310.45**p-value0.030.210.03N262126Enterprise is foreign-ownedCorrelation0.72***0.72***0.61***p-value0.000.000.01N211821Enterprise operates in international marketsCorrelation0.68**0.270.60***p-value0.000.320.01N201820Cooperated with partners which were located abroadCorrelation0.10-0.280.18p-value0.670.260.43N242024All correlations control for the natural log of the population size of each country, for cumulative gross fixed capital formation as a percentage of GDP, and for the surface area in square km. Innobarometer includes 29 countries and the Eurostat CIS4 database 27. N varies due to missing values in the internationalization indicators or control variables.
Significance level: * p<0.10; ** p<0.05; *** p<0.01
Appendix I: Description of the 29 indicators feeding into the European Innovation Scoreboard
ENABLERSHuman resources1.1.1 S&E and SSH graduates per 1000 population aged 20-29 (first stage of tertiary education) 1.1.2 S&E and SSH doctorate graduates per 1000 population aged 25-34 (second stage of tertiary education)1.1.3 Population with tertiary education per 100 population aged 25-641.1.4 Participation in life-long learning per 100 population aged 25-641.1.5 Youth education attainment level Finance and support1.2.1 Public R&D expenditures (% of GDP) 1.2.2 Venture capital (% of GDP) EVCA 1.2.3 Private credit (relative to GDP) 1.2.4 Broadband access by firms (% of firms) FIRM ACTIVITIESFirm investments2.1.1 Business R&D expenditures (% of GDP) 2.1.2 IT expenditures (% of GDP) EITO 2.1.3 Non-R&D innovation expenditures (% of turnover) Linkages & entrepreneurship2.2.1 SMEs innovating in-house (% of SMEs) 2.2.2 Innovative SMEs collaborating with others (% of SMEs) 2.2.3 Firm renewal (SME entries plus exits) (% of SMEs) 2.2.4 Public-private co-publications per million populationThroughputs2.3.1 EPO patents per million population 2.3.2 Community trademarks per million population OHIM 2.3.3 Community designs per million population OHIM 2.3.4 Technology Balance of Payments flows (% of GDP) OUTPUTSInnovators3.1.1 SMEs introducing product or process innovations (% of SMEs) 3.1.2 SMEs introducing marketing or organisational innovations (% of SMEs) 3.1.3 Resource efficiency innovators, unweighted average of: Share of innovators where innovation has significantly reduced labour costs (% of firms) Share of innovators where innovation has significantly reduced the use of materials and energy (% of firms) Economic effects3.2.1 Employment in medium-high & high-tech manufacturing (% of workforce) 3.2.2 Employment in knowledge-intensive services (% of workforce)3.2.3 Medium and high-tech manufacturing exports (% of total exports) 3.2.4 Knowledge-intensive services exports (% of total services exports) 3.2.5 New-to-market sales (% of turnover) 3.2.6 New-to-firm sales (% of turnover)Source: European Commission (2009)
Appendix II: Level B indicators are based on the following innovation-specific industry or product groups
B.1 FDI
Based on industries available from the OECDs International Direct investment Statistics we selected the following sectors:
Technology intensive manufacturing selected are: 2400 chemical products; 3000 office machinery and computers; 3200 radio, television, communication equipments. Other sectors have not been included due to a lack of data
Knowledge intensive services are: 7200 computer activities; 7300 research and development services; 7400 other business services
B.2 Trade
Based on products available from Eurostat (Science and Technology - High-tech industry and knowledge-intensive services) we derived trade in the following high-tech products:
Chemicals, computer-office machines, electrical machinery, electronics-telecommunications, non-electrical machinery, pharmacy, scientific instruments
With respect to trade in services we selected the following positions (credits and debits) from the IMFs Balance of Payments (IMF, 1993)
259. Computer and information services covers computer data and news-related service transactions between residents and non-residents. Included are data bases, such as development, storage, and on-line time series; data processingincluding tabulation, provision of processing services on a time-share or specific (hourly) basis, and management of facilities of others on a continuing basis; hardware consultancy; software implementationincluding design, development, and programming of customized systems; maintenance and repair of computers and peripheral equipment; news agency servicesincluding provision of news, photographs, and feature articles to the media; and direct, non-bulk subscriptions to newspapers and periodicals.
260. Royalties and license fees covers the exchange of payments and receipts between residents and non-residents for the authorized use of intangible, non-produced, nonfinancial assets and proprietary rights (such as patents, copyrights, trademarks, industrial processes, franchises, etc.) and with the use, through licensing agreements, of produced originals or prototypes (such as manuscripts and films). Inclusion of this item under services, rather than under income, is in accordance with the SNA treatment of such items as payments for production of services for intermediate consumption or receipts from sales of output used as intermediate inputs.
This paper is a development from the report Is the innovation performance of countries related to their internationalization? prepared for the European Commission in November 2009. The authors are grateful to the following people for offering useful comments on the original plan and/or on drafts of the report: Daniele Archibugi, Hugo Hollander, Keith Sequeira and Keith Smith. Some useful comments emerged also from the debate following the presentation of the report at the Workshop on Measuring Innovation: New Evidence in Support of Innovation Policy organized by Birkbeck, Merit and DG Enterprise and Industry, 29-30 October 2009, Birkbeck University of London. Most comments have been incorporated into this paper. Useful comments from three referees of this journal have led to significant improvement of the paper.
( Italian National Research Council, via Palestro, 32 00185 Rome, Italy, a.filippetti@irpps.cnr.it
( Birkbeck, University of London, Department of Management, Malet Street, London, WCIE 7HX, United Kingdom, m.frenz@bbk.ac.uk, corresponding author
( London South Bank University, Faculty of Business, 103 Borough Road, London, SE1 0AA, United Kingdom iettogg@lsbu.ac.uk
More recent support is in Amendola at al. (1993), Cantwell (1989, 1994), Cantwell and Sanna Randaccio (1993), Fagerberg, (1996) and Krugman (1995)
The problems of capturing the interactive process between innovation and internationalization elements (with respect specifically to R&D and exports) has been studied in Hughes (1986) and in Kleinknecht and Oostendorp (2002).
A detailed description of the methodology and variables feeding into the SII are available through the European Innovation Scoreboard (European Commission 2009a).
Details of industries and products included in the variables for level B are in Appendix II.
While the sources differ, all data for levels A and B originate with national statistical offices and are collected and published in accordance with international guidelines. Regarding level C, Innobarometer is designed and collected by Eurostat, and the CISs are collected by national statistical offices based on a harmonized questionnaire developed in collaboration with Eurostat. For both the Innobarometer and CIS we use the latest available waves of the surveys.
With respect to levels A and B, we computed correlations for each of the single sub-periods. The results are highly similar to those derived from the pooled data. We do not report them in the next section though they are available on request.
Data on gross fixed capital formation and surface area are taken from the World Development Indicators published by the World Bank.
Full results on the correlations without lags and with reverse lags are available upon request from the authors. Exploring associations based on reversed lags was suggested by one of the referees.
The products imported and exported may or may not belong to the same industrial category. The vast literature on intra-industry trade starting with Grubel and Lloyd (1975) is summarized in Grimwade (2000).
The authors concentrate on health sciences but suggest that their theory extends to other domains, particularly to the social sciences (p. 169).
PAGE
PAGE 1
Innovation
Internationalization
Globalization Index Level A
Globalization Index Level B
Globalization Index Level C
: ; Z [ \ ] _ d u v , 6 E L Q W n | C
G
ótmmimmmemm^ hAq hHA ha hx4 h9 hHA h9 hHA 6CJ aJ ! j #h9 hHA 0J 6CJ aJ ! j +h9 hHA 0J 6CJ aJ ! j *h9 hHA 0J 6CJ aJ h8 hHA 6CJ aJ mHsHh9 hHA 6hHA $j h9 hHA 0J 5CJ( UaJ( hHA 5CJ( aJ( h9 hHA 5CJ( aJ( h9 hHA 5 $ \ ] ^ _ ` a b c d a
$d a$gdHA d gdHA gdHA $a$gdHA U Z
4
a
v
` k ; = i m 9 : x f ~ h hHA nHtH j h9 hHA UnHtHh9 hHA nHtH h9 hHA hR 5h
h
h
h
5h? h9 hHA 5hrf hHA h9 hHA @
H : y y" z" {" " " &