Innovation and Economic Growth: An Empirical Investigation of European Countries

Transcription

1 Innovation and Economic Growth: An Empirical Investigation of European Countries Vetsikas, Apostolos (1); Stamboulis, Yeoryios (2); Markatou, Maria (1) 1: Department of Planning and Regional Development, University of Thessaly, Greece; 2: Department of Economics, University of Thessaly, Greece Abstract We examine the relationship between innovation outcomes and economic growth in a group of European countries for the period More specifically, three different types of intellectual property rights are used as indicators of innovation outcomes: patent applications, industrial design applications and trademark applications. Using the Johansen cointegration technique, we investigate possible long-term relationships between these three innovation output indicators and GDP per capita. We employ Granger causality analysis to check for causal relationships. The empirical findings show that the long-run relationships between innovation and economic growth are country and type of protection specific. The Granger causality test results indicate that patent applications have a stronger causal effect on economic growth rather than industrial design and trademark applications. The results show a unidirectional causal link from economic growth to intellectual property rights in Northern, mainly Nordic, European countries, while in Southern European countries the causal relationship is reverse. The implication of this study is that empirical analysis with emphasis on the combination of innovation output and input indicators may produce more insightful findings on the relationship between innovation and economic growth. Keywords: Innovation, Economic Growth, European Countries,, Granger Causality JEL codes: C22, O34, O40, O52

2 1. Introduction Europe maintains lofty ambitions for building its future growth and prosperity and safeguarding its social model through innovation (Veugelers et al., 2015). Innovation is far from being a recent phenomenon, and is inherent to human development. In recent years, the need to investigate and measure the relationship between innovation and economic growth has been expressed in the field of innovation economics. The use of econometric methodologies is an important tool to examine their relationship mainly at level of countries. In existing literature, research and development (R&D) expenditure (input indicator) and patents (output indicator) are widely used as a proxy for innovation partly because of the availability of data (Hassan and Tucci, 2010). Most researchers examine the possibility of cointegration and Granger causality relationship between innovation and economic growth for individual or multiple countries, but results are inconclusive. However, there is an evidence of long-term relationship in developed countries usually unidirectional causality relationship from economic growth to innovation. This means that a strong economy is a favorable condition for innovative activities, as developed countries have the opportunity to allocate financial resources in this direction. The aim for this research is to contribute to the investigation of the relationship between innovation and economic growth, with emphasis on innovation output indicators. Using the cointegration technique, we investigate possible long-term relationships between three innovation output indicators (patent applications, industrial design applications and trademark applications) and GDP per capita as a proxy of economic growth. Cullet (2005) argues that technology and knowledge, contained in these types of intellectual property, are important factors for economic growth and development and the contribution of technological innovation, with the protection of intellectual property rights, in economic growth is well established in the literature both theoretically and empirically (Nadiri, 1993; Gould and Gruben, 1996; Park, 2008). We employ Granger causality analysis to investigate causal relationships. The paper is organized as follows. We outline a review of relevant empirical studies in Section 2. We present the data used and the methodology of analysis in Section 3 and the empirical results in Section 4. In Section 5, we summarize, discuss and conclude. 2. Literature Review: R&D expenditure, Patents and Economic Growth The relationship between innovation and economic growth has emerged quite recently as a central and topical research theme in innovation economics. According to Maradana et al. (2017), studies on this subject may be categorized in four groups: Supply-leading hypothesis (SLH) assumes unidirectional causality from innovation activities to economic growth (see, for instance, Yang, 2006; Guloglu and Tekin, 2012; Cetin, 2013; Pradhan et al., 2016); [2]

3 Demand-following hypothesis (DFH) assumes unidirectional causality from economic growth to innovation activities (see, for instance, Sinha, 2008; Cetin, 2013; Sadraoui et al., 2014; Pradhan et al., 2016); Feedback hypothesis (FBH) assumes bidirectional causality between economic growth and innovation activities (see, for instance, Guloglu and Tekin, 2012; Cetin, 2013; Pradhan et al., 2016); Neutrality hypothesis (NLH)assumes no relationship between economic growth and innovation activities (see, for instance, Cetin, 2013; Pradhan et al., 2016) Modern econometric and statistical methods are used in order to investigate mostly the possibility of long-run relationship between growth (GDP per capita as the most commonly measure) and innovation (whether mainly measured by expenditure on R&D or patents data) both in developed and developing countries 2.1 Research and Development (R&D) expenditure and Economic Growth Literature Most empirical literature studies use R&D as measure of innovative activities. Sylwester (2001) examines the association between research and development (R&D) and the growth rate of output per capita at a national level in 20 OECD countries using a multivariate regression. The empirical findings show that there is not a strong association between the two but there is reported to be a positive relationship between industry R&D expenditure and economic growth. Ulku (2004) investigates the main postulations of the R&D based growth models that innovation is created in the R&D sectors and it enables sustainable economic growth. The analysis employs various panel data techniques and uses patent and R&D data for 20 OECD and 10 Non-OECD countries for the period The results suggest a positive relationship between per capita GDP and innovation in both OECD and non-oecd countries, while the effect of R&D stock on innovation is significant only in the OECD countries with large markets. Yu-Ming et al. (2007) investigate the cointegration and causal relationship between R&D expenditure and economic growth and examine the causality pattern in the R&D expenditure and economic growth in China from 1953 to The results suggest that there is a long-run cointegration relationship between the R&D and GDP, and a bidirectional causal relationship running from R&D to GDP and vice versa in the long-run also exists. Samimi and Alerasoul (2009) examine the impact of R&D on economic growth of 30 developing countries for the period The share of government expenditure on research in GDP, the number of researchers in each one million population and the scientific output of the countries are used as three different proxies for R&D. Their findings based on panel data regression models indicate that in general no significance positive impact exists in the countries under consideration. [3]

4 Cetin (2013) examines the causal relationship between R&D expenditure and economic growth, using the standard Granger (1969) and Toda-Yamamoto (1995) tests for causality, for 9 European countries for the period In consideration of standard Granger causality test, the empirical findings clearly exhibit that R&D expenditure cause GDP in the cases of Finland, France and Spain. The results also indicate that GDP causes R&D expenditure in Denmark and there is no causality between variables in other countries. On the other hand, the results of Toda-Yamamoto test imply that there is no causality between R&D expenditure and GDP in Holland, Ireland and Italy. However, there is bidirectional causality in Finland and France. The empirical findings also indicate that there is a causal relationship between variables running from R&D expenditure to GDP for Austria, while the direction of causality is from GDP to R&D expenditures for Denmark, Spain and Portugal. Consequently, this study provides further evidence supporting the hypothesis for some European countries. Sadraoui et al. (2014) investigate the Granger causality between R&D cooperation and economic growth in 32 industrial and developing countries from 1970 to They use an econometric method which is based on a panel test of the Granger non causality hypothesis. Using a new method to evaluate causality in a heterogeneous panel, they find that the causal relationship from R&D cooperation to economic growth is homogeneous among the panel. However, they find strong evidence of a heterogeneity of the causal relationship from economic growth to R&D cooperation in their sample. Santos and Catalao-Lopes (2014) investigate the causal relationship linking R&D and growth in a sample of 8 European Union (EU) countries, with an emphasis on Portugal. Specifically, they use annual OECD data for GDP and R&D, covering 22 observations, from 1987 to The empirical results, which are based on cointegration analysis, show that there is a stable long-run relationship between GDP and R&D only in the case of the United Kingdom. Moreover, they use Granger causality test to investigate the causal relationship linking R&D and growth in these countries. A causal relationship from growth to R&D can only be proven for France and Spain, whereas the inverse causality only seems to exist for the Netherlands. In addition, increased economic growth does not seem to necessarily mean increased R&D investment either. Pradhan et al. (2016), using a panel vector auto-regressive model, study interactions between innovation, financial development and economic growth in 18 Eurozone countries between 1961 and They focus on whether causality runs between these variables both ways, one way, the other way or not at all. The empirical results show that development of the financial sector and enhanced innovative capacity in the Eurozone contributes to long-term economic growth in the countries in the region. Empirical studies differ greatly in terms of level of analysis (companies, industries or countries), sources of data (time periods, countries) and measurements of key variables (stocks, flows or differences). Their results are not comparable; however, in general, the empirical results confirm theoretical assumptions that R&D expenditure has a positive and persistent effect on growth (Freimane and Balina, 2016). [4]

5 2.2 Patents and Economic Growth Literature In recent studies, researchers use patents as a measure of innovation activity to investigate the link between innovation and economic growth. In contrast with R&D expenditure, patents represent innovative output, indented to be commercialized (Hassan and Tucci, 2010). Sinha (2008) examines the relationship between patents and economic growth in Japan and South Korea using both individual country and panel data for The empirical findings show that for Japan the logarithms of real GDP and the number of patents are cointegrated. In addition, he finds a two-way causality between the growth of real GDP and growth of the number of patents in Japan. For South Korea, he does not find any evidence of cointegration and causality. For panel data, the logarithms of real and the number of patents are cointegrated. Panel causality tests find some evidence that the growth of real GDP Granger causes the growth of the number of patents. Hassan and Tucci (2010), using global patent data, empirically investigate the importance of both the quantity and quality of innovation on economic growth, controlling for past measures of inventive inputs. Moreover, their research examines how innovation inputs can be translated into per capita growth under the various economic structures and stages of economic development. The empirical results, based on a sample of 58 countries for the period , indicate that countries hosting firms with higher quality patents also have higher economic growth. Furthermore, there is some evidence that those countries that increase the level of patenting also witness a concomitant increase in economic growth. Saini and Jain (2011) examine the correlation between patent applications filed and financial growth of 9 selected Asian countries for the period of ten years ( ). The results concluded that it was a mixed result in case of Asian countries. Only, technology based countries economies were affected by patent applications filed. The countries having positive correlation (namely, Singapore, Thailand, Japan, Vietnam) depicts, leaving all other factors of affecting GDP, innovations are the major factor affecting GDP growth rate. Josheski and Koteski (2011) investigate the dynamic link between patent growth and GDP growth in G7 economies for the period Using ARDL methodology they show that there is a positive relationship in the long run between quarterly growth of patents and quarterly GDP growth. In the short run however at one or two lags there exist negative relationship between quarterly patents growth and quarterly growth of GDP. Granger causality test shows that patent growth Granger causes GDP growth in G7 countries and unrestricted VAR shows that there exists positive relationship between patent growth and GDP growth at two or three lags. Guzmán et al. (2012) examine the long-run relationship between economic activity in Mexico, measured by real GDP, and the number of patents granted to Mexican holders by the United States Patents and Trademark Office (USPTO) during the period Empirical [5]

6 evidence suggests that the marginal change in the patents affects the GDP growth rate but this does not have a significant effect on the number of patents change, that is, the number of patents is an exogenous variable. Moreover, the analysis of the impulse-response functions shows that shocks in the patents have length negative effects on real GDP, so as the effects of the real GDP shocks on the patents do. Maradana et al. (2017) examine the long-run relationship between innovation and per capita economic growth in the 19 European countries over the period They use six different indicators of innovation: patents-residents, patents-nonresidents, research and development expenditure, researchers in research and development activities, high-technology exports, and scientific and technical journal articles to examine this long-run relationship with per capita economic growth. Using cointegration technique, the study finds evidence of longrun relationship between innovation and per capita economic growth in most of the cases, typically with reference to the use of a particular innovation indicator. Using Granger causality test, the study finds the presence of both unidirectional and bidirectional causality between innovation and per capita economic growth. Their econometric results vary from country to country, depending upon the types of innovation indicators that they use in the empirical investigation process. The literature findings indicate that patents have impact on economic growth in developed countries. The causal effects vary from country to country, depending also to the choice of sample period. We have to stress that the existing literature is restricted for European countries, in the case of patent indicators and economic growth. The majority of studies use R&D expenditure as innovation proxy. In general, the literature empirical findings indicate the existence of a positive and strong impact of innovation on economic growth in developed countries. The cointegration methodology is widely used in empirical studies to identify long-term relationship between the variables under consideration, while the Granger causality test for possible causal relationships. The econometric results indicate the existence of long-term relationship in developed countries usually unidirectional causality relationship from economic growth to innovation. In this paper, we use innovation output indicators to capture this relationship. We examine the causal links between three types of IPRs and economic growth (in pairs) in a group of European countries, which constitutes the originality of this study. 3. Intellectual Property Rights, Data and Methodology We investigate the relationship between innovation and economic growth in a sample of European countries 1 for the period The period is selected according to availability of appropriate data. Here we use three different types of intellectual property as proxies of innovation activity, namely: 1 Austria, Belgium, Bulgaria, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, The Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom. [6]

7 Total patent applications (direct and PCT national phase entries), Total industrial design applications (direct and via the Hague system) and Total trademark applications (direct and via the Madrid system). A patent is a document, which contains structured and detail information regarding an invention and the probable application or use; it is accessible to the general public through national or international authorized agencies (Huang et al., 2003). The high inventive activity of patents is prominent among the other types of intellectual property rights (Jaffe et al., 1993; Jaffe and Trajtenberg, 1999). An industrial design is a document that describes in a concrete way products, services and systems, and as such links creativity to innovation (Hollanders and van Cruysen 2009). Moody (1980) originally highlighted the importance of the role of industrial design in technological innovation and Rothwell (1992) emphasized that industrial designs play a crucial role for successful industrial innovation. Despite the growing recognition of industrial design, few studies have attempted to quantify the contribution of good industrial design to company performance and economic growth (Hertenstein et al., 2005; Micheli and Gemser, 2016). Trademarks are words, signs, symbols of combination thereof that indentify goods and services as produced by a particular person or a company, therefore allowing consumers to distinguish between goods originating from different sources (Centi and Rubio, 2005). They play a crucial role in the process of marketing innovations, being instrumental in differentiating the attributes of goods and services in the marketplace (Mendonça et al., 2004). Trademarks rarely appear in public policy discussions as they are not considered to constitute particular problems for competition and innovation policy (Harhoff, 2006). The data are drawn from the database of the World Intellectual Property Organization (WIPO). We use the total applications for intellectual property rights to capture aggregate activity. GDP per capita is used as a proxy for growth (constant prices, $ 2010); it is derived from the database of the World Bank 2. Figure 1 illustrates the trends in total patent applications as to the total population in European countries under consideration. Norway presents by far the highest patent applications as to total population followed by Germany and United Kingdom. Belgium, Denmark, Ireland, Luxembourg decrease since the early 1990s. Southern European countries present a downward trend during the last decade except Italy where the decrease is less steep. 2 PAT denotes the number of applications for patents, DES, the number of applications for industrial designs and TR, the number of applications for trademarks. GDP per capita is symbolized by GDP. [7]

8 Figure 1: Trends in total patent applications as to total population 0,003 0,0025 0,002 0,0015 0,001 0, Austria Belgium Bulgaria Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Figure 2 shows the evolution of applications for industrial designs as to the total population 3. There is a remarkable growth trend in the cases of Switzerland and Norway during the last ten years and a notable decline in Austria since the early 2000s. In Southern European countries there is a remarkable increase during Germany, France and United Kingdom have a stable course for the period under consideration. Figure 2: Trends in total industrial design applications as to total population 0,001 0,0008 0,0006 0,0004 0, Austria Bulgaria Denmark Finland France Germany Greece Ireland Italy Norway Portugal Spain Sweden Switzerland United Kingdom 3 There is not available data for industrial design and trademark applications for Belgium, Luxembourg and the Netherlands. Also there is lack of data for industrial design and trademark applications in the case of Greece. [8]

9 Figure 3 illustrates the trends in total trademark applications as to the total population. During the last 20 years there has been a surge in national trademark applications in Europe. This increase in filings has been interpreted as a sign of increased innovative performance (Herz and Mejer, 2016). Switzerland has the highest growth rate followed by Norway. Figure 3: Total trademark applications as to total population 0,0045 0,004 0,0035 0,003 0,0025 0,002 0,0015 0,001 0, Austria Bulgaria Denmark Finland France Germany Greece Ireland Italy Norway Portugal Spain Sweden Switzerland United Kingdom Observing econometric methodology we first examine the stationary properties of the univariate time series. Stationarity tests in time series analysis help to avoid misleading results, as the problem of spurious regression (Granger and Newbold, 1974). Augmented Dickey-Fuller (ADF) test was used to test the unit roots of the concerned time series variables (Dickey and Fuller, 1979). For multivariate time series analysis involving stochastic trends, augmented Dickey Fuller (ADF) unit root tests were calculated for individual time series to provide evidence as to whether the variables are integrated. This was followed by multivariate cointegration analysis. If the hypothesis of a unit root is not rejected, then a test for cointegration is performed. The hypothesis tested is the null of non cointegration against the alternative of cointegration, using Johansen s maximum likelihood method. A vector autoregression approach was used to model each variable (which is assumed to be jointly endogenous) as a function of all the lagged endogenous variables in the system. We used the technique of cointegration according to the procedure of Johansen and Juselius (1990) to identify possible long-term economic relationships between the variables. In the Johansen framework, the first step is the estimation of an unrestricted, closed p-th order VAR in k variables. Johansen (1988) suggested two tests statistics to determine the cointegration rank. The first of these is known as the trace statistic: [9]

10 , (1) where, are the estimated eigenvalues λ 1 > λ 2 > λ 3 > >λ k and r 0 ranges from zero to k-1 depending upon the stage in the sequence. This is the relevant test statistics for the null hypothesis r r 0 against the alternative r r The second test statistic is the maximum eigenvalue test known as λ max ; we denote it as λ max (r 0 ). This is closely related to the trace statistic, but arises from changing the alternative hypothesis from r r 0 +1 to r = r The λ max test statistic is:, (2) The null hypothesis is that there are r cointegrating vectors, against the alternative of r + 1 cointegrating vectors. It is common practice to present the results of both tests while monitoring the trace is more powerful when considering two variables (Lutkepohl et al., 2001). We present the results of the trace test in Table 1. The basic assumption of this study is the possibility of strong long-term relationship in the case of patent applications and economic growth in the northern and more developed countries, while in the southern countries a long-run relationship is expected between the weaker intellectual property rights (industrial designs and trademarks) and GDP per capita. Also we assume that the impact of intellectual property rights on economic growth is more effective in upper middle countries compared to that in lower middle income countries (Janjua and Samad, 2007; Sattar and Mahmood, 2011). We employed Granger causality analysis to investigate causal relationships for one to six lags. According to Granger (1969), Y is said to Granger-cause X if and only if X is better predicted by using the past values of Y than by not doing so with the past values of X being used in either case. In short, if a scalar Y can help to forecast another scalar X, then we say that Y Granger-causes X. If Y causes X and X does not cause Y, it is said that unidirectional causality exists from Y to X. If Y does not cause X and X does not cause Y, then X and Y are statistically independent. If Y causes X and X causes Y, it is said that feedback exists between X and Y. Essentially, Granger s definition of causality is framed in terms of predictability. To implement the Granger test, a particular autoregressive lag length k (or p) is assumed and Models (3) and (4) are estimated by OLS: (3) (4) Furthermore, an F-test is carried out for the null hypothesis of no Granger causality; H 0 : b i1 b i2... b ik 0,i 1, where, the F statistic is the Wald statistic of the null hypothesis. If the F [10]

11 statistic is greater than a certain critical value for an F distribution, then we reject the null hypothesis that Y does not Granger-cause X, which means Y Granger-causes X. 4. Empirical Results All data are transformed into logarithmic returns in order to achieve mean-reverting relationships, and to make econometric testing procedures valid 4. The Augmented Dickey- Fuller test (ADF test) reveals that the hypothesis of a unit root test in LPAT, LDES, LTR and LGDP cannot be rejected even at the 5% significance level. The hypotheses of a unit root in DLPAT, DLDES, DLTR and DLGDP are rejected at least at the 5% level of confidence 5, indicating that all the variables in question are I(1). Table 1 summarizes the trace test results for cointegrating relationships between economic growth and each type of intellectual property rights 6. Table 1: Johansen Test Results Country LGDP - LPAT LGDP - LDES LGDP - LTR Austria Belgium n.a n.a Bulgaria Denmark Finland France Germany Italy Luxembourg n.a n.a The n.a n.a Netherlands Norway Portugal Sweden Switzerland United Kingdom 4 Table I (see Appendix) displays the estimates of the Augmented Dickey Fuller (ADF) test in levels and in first differences of the data with an intercept and trend. The ADF test results indicate that time series of LGDP are I(2) in the cases of Greece, Ireland and Spain. 5 LPAT is integrated of order one in Germany, Portugal and United Kingdom at the 10% level of confidence. Moreover, LGDP is integrated of order one in Bulgaria and Norway at the 10% level of confidence. 6 The analytic trace test results are presented in Table II to IV (in the Appendix). Variables LPAT, LDES, LTR, LGDP, Maximum lag in VAR=2. [11]

12 The findings show that there are long run relationships between GDP per capita and patent applications in the cases of Austria, Denmark, France, Germany, Italy, the Netherlands, Sweden and United Kingdom at the 5% level of confidence. In Belgium there is a long run relationship at the 10% level of confidence. In the case of GDP per capita and industrial design applications, there are long run relationships in Austria, Finland, France, Italy, Portugal, Sweden and United Kingdom; in Italy and Sweden results are more significant. The last column (Table 1) indicates long run relations between GDP per capita and trademark applications in Austria, Finland, France, Italy, Norway, Portugal and Switzerland; the findings are more significant in Austria, Italy and Norway. Consequently, the econometric results vary from country to country, depending upon the types of innovation indicators that we use in the empirical investigation process. In Table 2 we summarized the results of the Granger causality test between GDP per capita and patents applications 7. Table 2: Granger Causality Test Results between LPAT and LGDP Direction of Causality LPAT LGDP Lags Country Austria ** * Belgium *** * Bulgaria ** ** ** Denmark * Finland * ** Italy ** ** Luxembourg ** The Netherlands ** ** ** * Switzerland * * LGDP LPAT Bulgaria *** * Germany *** * ** *** *** *** Norway * ** * ** * ** Portugal * Sweden *** ** ** ** ** * Switzerland * * ** United Kingdom * LGDP LPAT France *, **, *** denote significance at 10%, 5% and 1% respectively. This note also applies to the subsequent tables. There is a unidirectional causal link from patent applications to GDP (LPAT LGDP) in Austria, Belgium, Denmark, Finland, Italy, Luxembourg and the Netherlands. A reverse causal relationship (LGDP LPAT) is indicated in Germany, Norway, Portugal, Sweden and United Kingdom; the causal links are stronger in Germany, Norway and Sweden. 7 The analytic Granger causality test results for the three pairs of variables are presented in Table V to VII (see Appendix). [12]

13 Bidirectional relations are indicated in the cases of Bulgaria and Switzerland (LGDP LPAT). In France, there is no evidence of bidirectional or unidirectional causal links between patent applications and GDP per capita. Table 3 reports the causal relationships between GDP per capita and industrial design applications. Table 3: Granger Causality Test Results between LDES and LGDP Direction of Causality LDES LGDP Lags Country Austria * France *** Italy ** United Kingdom * LGDP LDES Austria * Denmark *** * ** Finland * *** ** ** * France ** Portugal * ** ** * Sweden *** ** *** *** *** *** United Kingdom ** LGDP LDES Bulgaria Germany Norway Switzerland The empirical findings indicate a unidirectional causal link from industrial design applications to GDP (LDES LGDP) in Italy for one lag. A reverse causal link (LGDP LDES) is found in Denmark, Finland, Portugal and Sweden. Bidirectional causal relations are indicated in Austria, France and United Kingdom but only for one lag. Regarding the rest four European countries (Bulgaria, Germany, Norway and Switzerland) Granger causality test finds no evidence of causal links. Table 4 presents the causal relationships between GDP per capita and trademark applications. Table 4: Granger Causality Test Results between LTR and LGDP Direction of Causality LTR LGDP Lags Country Austria ** France ** Italy * Portugal *** LGDP LTR Austria ** *** ** *** Denmark *** *** ** Finland ** *** ** ** ** ** Norway * ** * * [13]

14 Sweden ** * * LGDP LTR Bulgaria Germany Switzerland United Kingdom Table 4 indicates that there is a unidirectional causal link from trademark applications to GDP (LTR LGDP) in France, Italy and Portugal for one lag. The relationships between LGDP and LTR for the four Nordic countries are unidirectional running from LGDP to LTR. In Austria the causal link is bidirectional but is stronger from the variable of economic growth to trademark applications. Regarding the rest four European countries (Bulgaria, Germany, Switzerland and United Kingdom) Granger causality test finds no evidence of causal links. In Table 5 we summarize the Granger causality test results. Table 5: Summarized causalities (time lags in parentheses) Direction of Causality From IPR type to GDP Intellectual Property Rights Type Patents Industrial Designs Trademarks Austria (1,3) Belgium (1,2) Bulgaria (4,5,6) Denmark (3) Finland (1,2) Italy (1,2) Luxembourg (5) The Netherlands (1,3,4) Switzerland (1,2) Austria (1) France (1) Italy (1) United Kingdom (1) Austria (1) France (1) Italy (1) Portugal (1) From GDP to IPR type Bulgaria (5,6) Germany (1,2,3,4,5,6) Norway (1,2,3,4,5,6) Portugal (6) Sweden (1,2,3,4,5,6) Switzerland (4,5,6) United Kingdom (5) Austria (1) Denmark (1,2,6) Finland (2,3,4,5,6) France (1) Portugal (1,2,3,4) Sweden (1,2,3,4,5,6) United Kingdom (1) Austria (2,3,4,5) Denmark (1,2,3) Finland (1,2,3,4,5,6) Norway (1,2,3,4) Sweden (1,2,5) No causality Bulgaria Germany France Norway Switzerland Bold characters denote significance at 5% or 1%. Bulgaria Germany Switzerland United Kingdom The main econometric results may be summarized as follows: The cointegration approach shows that the long-run relationship between economic growth and intellectual property rights is country and type of [14]

15 protection specific. The mixed results confirm empirical findings of recent studies which investigate the possibility of cointegration between innovation and economic growth for European countries (Santos and Catalao-Lopes, 2014; Maradana et al., 2017). Patent applications have a stronger effect on economic growth rather than industrial design and trademark applications. The economic effects of industrial design and trademark applications are immediate in contrast to patent applications where their effects become visible after at least two lags in several countries under consideration. Probably, certain sectors where patents are more significant and common (e.g. pharmaceuticals, nanotechnology, etc.) are relatively stronger in these countries, hence they have causal effects on growth in the long run. In general, there is a powerful causal relationship running from economic growth to three different types of innovation for a significant number of time lags. The results are quite distinct in the cases of Austria, Denmark, Finland, Norway and Sweden. This means that a strong economy may provide more incentives for innovation activity and for vesting of intellectual property rights. There is evidence for a different direction of causality between Nordic (Denmark, Finland, Norway and Sweden plus Germany which fits the pattern only in the case of patents) and Southern Latin European countries (mainly Italy and France). Specifically, the empirical results indicate a one-way causal relationship from economic growth to intellectual property rights in Northern European countries. These countries are able to allocate more resources to R&D expenditure. On the contrary, in Southern European countries and mainly in Italy, there is evidence of reverse causal link, from intellectual property rights to economic growth, which requires further analysis. However, the problems which were identified in the time series of GDP (in logarithms) in Greece and Spain restrict the generalization of empirical findings and possible differences between Northern and Southern European countries. It is important to highlight that results may be sensitive to the choice of sample period, selection of variables and methodology adopted. This also indicates the sensitivity of Granger causality and that is why results based on Granger causality should be interpreted with care. 5. Concluding Remarks The relationship between innovation and economic growth has emerged quite recently as a central and topical research theme in innovation economics. The empirical studies investigate mostly the possibility of long-run relationship between innovation (whether mainly measured by expenditure on R&D or patents data) and economic growth. The majority of literature [15]

16 studies use R&D as measure of innovative activities and the results indicate that R&D expenditure has a positive and persistent effect on economic growth. However, in recent studies, researchers use patents as a measure of innovation activity. The empirical findings indicate the existence of long-term relationship in developed countries usually unidirectional causality relationship from economic growth to innovation. In this paper, we examined the relationship between innovation and economic growth in a group of European countries for the period We used three different types of intellectual property rights to capture innovation activity output: patent applications, industrial design applications and trademark applications. The empirical results reported herein suggest that: The long-run relationship between economic growth and intellectual property rights is country and type of protection specific. Among the three types of IPRs, patent applications have a stronger causal effect on economic growth than industrial design and trademark applications. There is a strong causal relationship running from economic growth to all types of IPRs for a significant number of time lags, which indicates that a strong economy is a favorable condition for IPR activity. There is evidence for a different direction of causality between Northern (especially Nordic) and Southern Latin European countries. In the first ones, the causal relationship seems to be from economic growth to innovation output indicators while in Southern countries the causal link is reverse. The empirical studies focusing on R&D expenditure as innovation proxy indicate a unidirectional causal relationship from innovation indicator to economic growth in developed countries. In this paper, the causal links seem to be different in the case of Southern European countries. However, there is a shortage in literature for IPRs and economic growth for European countries. As a result, the comparison with literature studies cannot be accurate. The empirical findings and possible differences between Northern and Southern European countries create the need for discussion about measures of innovation activity (Edquist and Zabala-Iturriagagoitia, 2015). The European Innovation Scoreboard (EIS) places Northern European countries (mainly Sweden, Denmark and Finland) on top, as innovation leaders in the most recent reports. These countries, by-and-large due to their strong economy, are able to allocate significant resources to R&D expenditure (public and business). The statistical information provided by the EIS and the results of their analysis based on it should be supplemented with other more contextual and qualitative information regarding the innovation system under study (Foray and Hollanders, 2015). The approaches with input indicators (e.g. R&D expenditure, number of researchers in R&D, venture capital as % of GDP etc) as innovation proxies are likely to generate misleading results about the countries' innovation performance, as they do not take into account the innovation performance indicated from the transformation of inputs to outputs (Edquist and Zabala-Iturriagagoitia, 2015). The innovation performance of an innovation system should be understood from two perspectives: (i) the delivery-production of innovation outputs; and (ii) the innovation [16]

17 efficiency of the system as a whole (Edquist and Zabala-Iturriagagoitia, 2015). Alternative methodologies with clear separation of inputs and outputs can provide a completely different picture (Havas, 2015). Southern European countries significantly fall behind Europe 2020 goals (in terms of R&D expenditure). These countries, which are described as moderate innovators (European Union, 2013; 2014; 2015; 2016) may have higher performance ratios (output/input) in terms of innovation (Edquist and Zabala-Iturriagagoitia, 2015). Still, in qualitative terms, which are not and possibly cannot be derived from this type of quantitative analysis, their innovation activity may be lacking in strategic significance. In future research, innovation indicators available from the European Innovation Scoreboard (EIS) could be exploited by using panel data techniques. A separation of European countries in two or more groups (e.g. Northern and Southern European countries) may provide quite different results about the efficiency of individual national innovation systems. The use of panel data techniques and the recent panel causality tests of Dumitrescu and Hurlin (2012), Tervo (2009) and Hartwig (2010) could be applied to investigate possible causal relationships between innovation indicators and economic growth. The econometric approaches in the literature, as presented in Section 2, use either R&D expenditure or patents as innovation indicators and therefore they fail to capture the full range of innovation activities and the sectoral expertise in each country. Possible broadening of the range of (structural) indicators explored could yield a richer picture. Further empirical analysis with emphasis on the combination of innovation output and input indicators may produce more insightful findings on the relationship between innovation and economic growth. [17]

18 References Akcay, S. (2011), Causality Relationship Between Total R&D Investment and Economic Growth: Evidence from United States, Journal of Faculty of Economics and Administrative Sciences, 16 (1), pp Centi, J. and Rubio, N. (2005), Intellectual property rights and entrepreneurship: On the precedence of trademarks, International Centre for Economic Research Working Paper Series, No.29. Cetin, M. (2013), The hypothesis of innovation-based economic growth: a causal relationship, International Journal of Economic and Administrative Studies, 6 (11), pp Cullet, P. (2005), Intellectual Property Protection and Sustainable Development, New Delhi: LexisNexis Butterworths. Dickey, D. and Fuller, W. (1979), Distribution of the Estimators for Autoregressive Series with a Unit Root, Journal of the American Statistical Association, 74, pp Dumitrescu, E.I. and Hurlin, C. (2012), Testing for Granger non-causality in heterogeneous panels, Economic Modelling, 29 (4), pp Edquist, C. and Zabala-Iturriagagoitia, J.M. (2015), The Innovation Union Scoreboard is Flawed: The case of Sweden not being the innovation leader of the EU, CIRCLE Papers in Innovation Studies, No. 2015/16. European Union (2013), Innovation Union Scoreboard 2013, Brussels: European Commission. European Union (2014), Innovation Union Scoreboard 2014, Brussels: European Commission. European Union (2015), Innovation Union Scoreboard 2015, Brussels: European Commission. European Union (2016), European Innovation Scoreboard 2016, Brussels: European Commission. Foray, D. and Hollanders, H. (2015), An assessment of the Innovation Union Scoreboard as a tool to analyse national innovation capacities: The case of Switzerland, Research Evaluation, 24, pp Freimane, R. and Bāliņa, S. (2016), Research and Development Expenditures and Economic Growth in the EU: A Panel Data Analysis, Economics and Business, 29 (1), pp [18]

19 Gould, D. M. and Gruben, W. C. (1996), The Role of Intellectual Property Rights in Economic growth, Journal of Development Economics, 48, pp Granger, C. W. J. (1969), Investigating Causal Relation by Econometric and Cross- Sectional Method, Econometrica, 37 (3), pp Granger, C. W. J. and Newbold, P. (1974), Spurious regressions in econometrics, Journal of Econometrics, 2, pp Guloglu, B. and Tekin, B. (2012), A panel causality analysis of the relationship among research and development, innovation, and economic growth in highincome OECD countries, Eurasian Economic Review, 2 (1), pp Gumus, E. and Celikay, F. (2015), R&D Expenditure and Economic Growth: New Empirical Evidence, The Journal of Applied Economic Research, 9 (3), pp Guo, Y. and Wang, B. (2013), Study on the Economic Growth of Patent Output in the High-tech Industry, Journal of Management and Sustainability, 3 (1), pp Guzman, C.A., Lopez-Herrera, F. and Venegas-Martinez, F. (2012), A Analysis between Patents and Economic Growth in Mexico, , Investigación económica, 71, pp Harhoff, D. (2006), Intellectual property rights in Europe where do we stand and where should we go?, A contribution to the project Globalisation Challenges for Europe and Finland organized by the Secretariat of the Economic Council, Prime Minister s Office, Economic Council of Finland. Hartwig, J. (2010), Is health capital formation good for long-term economic growth? Panel Granger-causality evidence for OECD countries, Journal of Macroeconomics, 32, pp Hassan, I. and Tucci, C. L. (2010), The innovation-economic growth nexus: global evidence, Research Policy, 39 (10), pp Havas, A. (2015), Various approaches to measuring business innovation: their relevance for capturing social innovation, Institute of Economics CERS HAS Discussion Papers, MT-DP 2015/54. Hertenstein, J., Platt, M., and Veryzer, R. (2005), The Impact of Industrial Design Effectiveness on Corporate Financial Performance, Journal of Product Innovation Management, 22, pp Herz, B. and Mejer, M. (2016), On the fee elasticity of the demand for trademarks in Europe, Oxford Economic Papers, 68 (4), pp [19]

20 Hollanders, H. and van Cruysen, A. (2009), Design, Creativity and Innovation: A Scoreboard Approach. PRO INNO EUROPE. Huang, Z., Chen, H., Yip, A., Ng, G., Guo, F. and Chen, Z.K. (2003), Longitudinal patent analysis for nanoscale science and engineering: Country, institution and technology field, Journal of Nanoparticle Research, 5, pp Jaffe, A. B. and Trajtenberg, M. (1999), International knowledge flows: evidence from patent citations, Economics of Innovation and New Technology, 8 (1/2), pp Jaffe, A. B., Trajtenberg, M. and Henderson, R. (1993), Geographic localization of knowledge spillovers as evidenced by patent citations, Quarterly Journal of Economics, 108 (3), pp Janjua, Z. P. and Samad, G. (2007), Intellectual property rights and economic growth: the case of middle income developing countries, The Pakistan Development Review, 46 (4), pp Johansen, S. (1988), Statistical Analysis of Co-integration Vectors, Journal of Economic Dynamics and Control, 2, pp Johansen, S. and Juselius, K. (1990), Maximum likelihood estimation and inferences on cointegration: With Applications to the Demand for Money, Oxford Bulletin of Economics and Statistics, 52 (2), pp Josheski, D. and Koteski, C. (2011), The Causal Relationship Between Patent Growth and Growth of GDP with Quarterly Data in the G7 Countries:, ARDL and Error Correction Models, MPRA Paper No Lutkepohl, H., Saikkonen, P. and Trenkler, C. (2001), Maximum eigenvalue versus trace tests for the cointegrating rank of a VAR process, Econometrics Journal, 4, pp Maradana, R., Pradhan, R., Dash, S., Gaurav, K., Jayakumar, M. and Chatterjee, D. (2017), Does innovation promote economic growth? Evidence from European countries, Journal of Innovation and Entrepreneurship, 6 (1), pp Mendonca, S., Pereira, T. and Godinho, M. (2004), Trademarks as an indicator of innovation and industrial change, Research Policy, 33 (9), pp Micheli, P. And Gemser, G. (2016), Signaling strategies for innovative design: A study on design tradition and expert attention, The Journal of Product Innovation Management, 33 (5), pp Moody, S. (1980), The role of industrial design in technological innovation, Design Studies, 1 (6), pp [20]

21 Nadiri, M. I. (1993), Innovations and Technological Spillovers, Working Paper 4423, National Bureau of Economic Research, Cambridge, MA. Park, W.G. (2008), Intellectual property rights and international innovation, In: Keith Maskus (Ed.), Frontiers of Economics and Globalization, 1, Handbook Series, Elsevier, Amsterdam Netherlands, pp Pradhan, R. P., Arvin, M. B., Hall, J. H., and Nair, M. (2016), Innovation, financial development and economic growth in eurozone countries, Applied Economics Letters, 23 (16), pp Rothwell, R. (1992), Successful industrial innovation: Critical factors for the 1990s, R&D Management, 22 (3), pp Sadraoui, T., Ali, T. and Deguachi, B. (2014), Economic growth and international R&D cooperation: a panel granger causality analysis, International Journal of Econometrics and Financial Management, 2 (1), pp Saini, A. and Jain, S. (2011), The Impact of Patent Applications Filed on Sustainable Development of Selected Asian Countries, BVICAM s International Journal of Information Technology, 3 (2), pp Samimi, A.J. and Alerasoul, S.M. (2009), R&D and Economic Growth: New Evidence from Some Developing Countries, Australian Journal of Basic & Applied Sciences, 3 (4), pp Santos, J. F. and Catalao-Lopes, M. (2014), Does R&D matter for economic growth or vice-versa? An application to Portugal and other European countries, Archives of Business Research, 2 (3), pp Sattar, A. and Mahmood, T. (2011), Intellectual property rights and Economic growth: Evidences from high, middle and low income countries, Pakistan Economic and Social Review, 49 (2), pp Sinha, D. (2008), Patents, innovations and economic growth in Japan and South Korea: evidence from individual country and panel data, Applied Econometrics and International Development, 8 (1), pp Sylwester, (2001), R&D and Economic Growth, Knowledge, Technology, & Policy, 13 (4), pp Tervo, H. (2009), Centres and peripheries in Finland: Granger causality tests using panel data, Spatial Economic Analysis, 4, pp Toda, H.Y. and Yamamoto, T. (1995), Statistical Inference in Vector Autoregressions with Possibly Integrated Processes, Journal of Econometrics, 66, pp [21]