Netherlands Observatory of Science and Technology. Science and Technology Indicators Summary

2008 Netherlands Observatory of Science and Technology Science and Technology Indicators Summary

Science and Technology Indicators 2008 Summary Netherlands Observatory of Science and Technology NOWT

This summary report is based on NOWT s Dutch-language report Wetenschaps- en Technologie- Indicatoren Rapport 2008 (Science and Technology Indicators Report 2008) that was released in March 2008. For more information about NOWT, or NOWT s series of Science and Technology Indicators Reports, please visit the NOWT website (www.nowt.nl) or contact: Robert Tijssen (NOWT coordinator) Centre for Science and Technology Studies (CWTS) Leiden University, The Netherlands Tel: +31 71 5273960 Fax: +31 71 5273911 E-mail: tijssen@cwts.leidenuniv.nl For complementary copies of this summary report, or the Dutch-language full report, please contact: Jan van Steen Ministry of Education, Culture and Science Directorate Research and Science Policy PO Box 16375, 2500 BJ Den Haag, The Netherlands Tel: +31 70 4123756 Fax: +31 70 4122080 E-mail: j.c.g.vansteen@minocw.nl This summary report and the Dutch-language full version are electronically available at www.nowt.nl

2008 Science and Technology Indicators Summary Netherlands Observatory of Science and Technology (NOWT) NOWT is a formal collaboration between Centre for Science and Technology Studies (CWTS), Leiden University, The Netherlands Maastricht Economic and social Research and training centre on Innovation and Technology (UNU-MERIT), Maastricht, The Netherlands Robert Tijssen Hugo Hollanders Thed van Leeuwen Anton Nederhof CWTS UNU-MERIT CWTS CWTS www.nowt.nl

Table of contents 1 Introduction 6 1.1 Science, technology and innovation policy framework in the Netherlands 6 1.2 Quantitative analysis and benchmarking of the Dutch R&D system 7 2 R&D resources 10 2.1 R&D expenditures 10 2.2 Human resources devoted to R&D 17 3 R&D performance 18 3.1 Research publication output 18 3.2 Patent output 24 4 R&D co-operation 28 4.1 Co-operation between industry and the public sector 28 4.2 Research co-operation 31 5 General conclusions 32 List of abbreviations 34 Acknowledgements 35 5

1Introduction 1.1 Science, technology and innovation policy framework in the Netherlands The Netherlands is a relatively small country located in the North-Western part of Europe. With a population of 16.4 million it is one of the most densely populated areas on our globe. It is one of the EU-27 member states and was one of the founders of the EU in 1950. The Netherlands can boast modern, high-tech infrastructures and public facilities. It is among the world s leading countries in terms of economic performance and it is the home base of several of Europe s largest companies. Among the most important contributing factors of the open Dutch economy are its export position and its innovative capacity. The services sector accounts for more than 50% of its GDP. Industrial firms and the services-oriented business firms are increasingly internationally oriented; they are competing on the European and global markets. The advanced economy of the Netherlands is gradually becoming more knowledge-intensive, where economic growth and prosperity is increasingly based on a highly skilled labour force and innovative products, services and processes. A sustainable knowledge-intensive economy requires an excellent educational system, a broad high-quality science base, and a competitive R&D system. The national government recognizes that education and research are vital for the further development of this knowledge-based society. The Netherlands has embraced the EU s Lisbon and Barcelona objectives to improve the competitive performance of Europe. Government policy for research and innovation at the national level is the joint responsibility of the Ministry of Economic Affairs (trade, industry, technology and innovation) and the Ministry of Education, Culture and Science. At the political and policy level, the inter-departmental Committee on Economy, Knowledge and Innovation prepares the proposals in the field of research for discussion in the Council on Economy, Knowledge and Innovation, and, subsequently, in the Ministerial Council itself. The national government has also established an Innovation Platform, presided by the Prime Minister, to aid in formulating a coherent strategy on research and innovation. A wide range of policy actions and instruments have been initiated, or are currently being planned, by government agencies to further the role of science, technology and innovation in the Netherlands. 1 Science and innovation policy-making by the Dutch government has a long tradition of using statistical data for assess- 6

ing the state of the R&D system in the Netherlands. Policy making is becoming increasingly evidence-based, fed by systemic overviews and supported by reliable information sources and series of internationally recognized quantitative indicators. The Netherlands Observatory of Science and Technology (denoted here by its Dutch acronym NOWT) is one of the main government sources of factual information. NOWT s Science and Technology Indicators Report series, issued by the Ministry of Education, Culture and Science, deals with the overall performance of the R&D system, thus cutting across key policy areas of both ministries responsible for policy on research and innovation. 2,3 However, the main focus of the report is on the Dutch science base, and its interrelations with the higher education system and research-intensive industry, within the broader framework of the Dutch R&D system. The aim of Dutch science policy is to create a research climate that fosters world-class scientific achievements and promotes the welfare and well-being of society at large. Specific science policy measures deal with expanding the scope and resources for fundamental research, introduce strategies to promote innovation and the utilization of research results, strengthen the self-regulating powers of the scientific community, and improve career opportunities for young researchers and women. 1.2 Quantitative analysis and benchmarking of the Dutch R&D system of R&D-performing institutional sectors and individual universities, research institutes and large R&D-intensive companies. The dominant type of data is secondary data from the OECD, EC, or other official sources. The bibliometric data are generated by CWTS and describe output-based key features of the Dutch research system. The nature of the analyses in this report differs per theme, but the emphasis is on quantitative indicators and analysis. The international frame of reference comprises the following countries: Australia (AUS), Austria (AUT), Belgium (BEL), Canada (CAN), Denmark (DNK), Finland (FIN), France (FRA), Germany (GER), Ireland (IRE), Japan (JPN), Republic of Korea (KOR), the Netherlands (NLD), Norway (NOR), Sweden (SWE), Switzerland (CHE), United Kingdom (UK), and the United States (US). Each of these OECD member states is among the most advanced countries worldwide in terms of R&D performance, innovation and economic competitiveness. The overview of the summary findings is displayed in Table 1. The Netherlands shows an excellent record in knowledge creation, with a highly productive science base. However we find comparative weaknesses in R&D intensity, in particular business R&D. Overall, the Netherlands ranks around the average for the benchmark countries according to the International Competitiveness Index. For purposes of policy debate and design, but also for public accountability, government agencies like the Dutch Ministry of Education, Culture and Science need to monitor, as accurately as possible, structural changes and trends in scientific and technological performance - not only from the national perspective, but increasingly also within a European or global comparative framework. In line with previous editions, the 2008 version of NOWT s S&T indicators report provides an indicator-based scan of the performance of the Netherlands within such an international comparative perspective. It assembles an array of empirical evidence from domestic sources and international statistical agencies to analyze and monitor national and international trends within R&D systems. The five analytical chapters of the full Dutch-language report therefore comprise of: (1) a variety of macro-level analyses based on country-level comparisons, and (2) several meso-level analyses detailing the state of affairs in the Netherlands. These analyses are conducted both at the level 1 These initiatives include fiscal measures to stimulate R&D, FESfunds for public-private R&D projects in strategic areas, and human resources targeted measures. See the following policy documents: Ministry of Economic Affairs and Ministry of Education, Culture and Science, Science, Technology and Innovation in the Netherlands: policies, facts and figures 2006, 2006; Ministry of Education, Culture and Science, Strategic agenda for higher education, research and science policy, 2007; Innovation Platform, Kennisinvesteringsagenda 2006-2010, 2006. 2 These S&T Indicators reports are published bi- or tri-annually and present a wide range of quantitative indicators and statistical data. Other recent government-issued reports (in Dutch) comprising statistics on the Dutch R&D and innovation system are: Kennis in kaart 2007 (Ministry of Education, Culture and Science, 2007), Kennis en Economie 2007 (Statistics Netherlands, 2007), and Het Nederlandse ondernemingsklimaat in cijfers 2007 (Statistics Netherlands, 2007). 3 The Ministry of Economic Affairs is responsible for innovation policy. 7

Table 1. The Netherlands: competitive performance at relatively low costs Summary performance indicators for the Netherlands and benchmark countries (2000-2007) US CHE DNK SWE DNK FIN JPN UK NLD KOR CAN AUT NOR FRA AUS BEL IRE International competitive 1 2 3 4 5 6 8 9 10 11 13 15 16 18 19 20 22 ranking Productivity scientific research 15 1 9 6 13 14 17 4 2 16 8 10 11 12 5 7 3 Productivity technological 6 1 9 4 5 2 3 12 7 17 14 11 13 10 15 8 16 development % R&D expenditure total 5 4 8 1 7 2 3 13 14 6 11 10 16 9 15 12 17 % R&D expenditure private 5 4 9 2 8 6 3 12 14 1 15 11 16 13 17 7 10 sector % R&D expenditure higher 16 7 8 11 14 13 15 9 2 17 1 4 3 12 5 10 6 education % R&D expenditure research 7 17 12 16 5 10 11 8 4 6 9 15 3 2 1 14 13 institutes * Performance indicators and measurements: International competitive ranking Ranking in Global Competitiveness Index (2007) Productivity scientific research Research publications per capita R&D staff (2002-2004) Productivity technological development Triadic patents per capita R&D staff (2000-2005) % R&D expenditure - total Total R&D expenditure as % GDP (2000-2005) % R&D expenditure private sector R&D expenditure by industry as % GDP (2000-2005) % R&D expenditure - higher education R&D expenditure by higher education as % GDP (2000-2005) % R&D expenditure research institutes R&D expenditure by research institutes as % GDP (2000-2005) Sources: OECD MSTI database; World Economic Forum, Global Competitiveness Index, Thomson Scientific/CWTS Web of Science database. Data treatments: UNU-MERIT and CWTS. This performance profile seems to represent a paradox between a relatively weak performance in R&D expenditure and a strong performance in scientific and technological output. The next chapters will shed more light on this state of affairs by providing background information and relevant statistics about the R&D performance of the Netherlands within an international comparative framework. The final chapter summarizes key results and draws general conclusions. It should be stressed that the analysis in the next chapters does not fully reflect the level of detail in the Dutch-language report, which provides a wider range of international benchmark analyses alongside more detailed information on the different actors within the Dutch R&D system and knowledge production of each of the Dutch universities. Readers are encouraged to consult the NOWT website (www.nowt.nl) where all the graphs and tables from the Dutch-language report are separately available. 8

2R&D resources 2.1 R&D expenditures Investments in science and new technologies are measured by the amount of expenditures in R&D (Research and Development). Although R&D does not capture all scientific and technological activities, it is nonetheless considered to be one of the leading indicators for benchmarking the R&D performance of countries within an international setting. R&D intensities, defined as the percentage share of R&D in a country s GDP (Gross Domestic Product), are used by national governments, the OECD, EC and Eurostat to evaluate the relative size of new investments in research activities and to benchmark countries within Europe and worldwide. 4 The Netherlands spent about 8.8 billion euros on R&D in 2005 5. The relative level of spending in the Netherlands, measured by R&D intensity, is 1.73% of GDP. The Dutch R&D intensity is not only below the average of the benchmark countries, it is also below that of the EU25 (Figure 2). The Netherlands is lagging behind most countries, although Ireland and Norway have an even lower R&D intensity. In real terms, i.e. corrected for inflation, Dutch R&D expenditures have hardly grown over the past 5 years. With stable R&D expenditures and a growing GDP (2% per year in real terms), we can clearly see why the R&D intensity has fallen significantly compared to the 1.96% in 1999. The Dutch growth rate of real R&D expenditures is much lower than that in most other countries; Figure 2 shows that the Netherlands is showing the overall worst performance and is in clear danger of falling behind its main competitors. Real R&D expenditures have grown fastest in South Korea, Ireland, Australia and Austria, and we can observe a significant increase in R&D intensities in South- Korea, Australia, Austria and Switzerland. On the other hand, R&D intensities have fallen in Belgium, France, Norway, Sweden, the UK and the Netherlands. 4 Innovation performance of most European countries and some non- European OECD countries is measured annually by the European Innovation Scoreboard (http://www.proinno-europe.eu/metrics). In this scoreboard the Netherlands is performing above the EU average performance, but is lagging well behind Sweden, Switzerland and Finland, the so-called innovation leaders. 5 Total R&D expenditures are also referred to as Gross Domestic Expenditure on R&D or GERD. 10

Figure 2. The Netherlands is falling behind the most advanced countries in terms of R&D expenditure Total R&D intensities (2005) and total R&D 5-year growth rates (1999-2005)* Figure 3. The Netherlands is falling behind due to low growth rates in private sector R&D expenditure Private sector R&D intensities (2005) and 5-year R&D growth rates (1999-2005)* R&D intensity(%) 4,5 FOLLOWERS Average for benchmark countries LEADERS R&D-intensity (%) 3,5 FOLLOWERS Average for benchmark countries LEADERS 4,0 3,5 3,0 2,5 2,0 1,5 1,0 Average for benchmark countries NLD GER FRA BEL SWE UK US EU25 FIN JPN CHE DNK CAN NOR AUT AUS IRE KOR 3,0 2,5 2,0 1,5 1,0 0,5 SWE US GER FRA BEL UK EU25 CAN NLD NOR JPN FIN CHE DNK IRE AUT AUS KOR Average for benchmark countries FALLING BEHIND CATCHING-UP 0,5-2 0 2 4 6 8 10 Average annual growth rate real R&D expenditures(%) FALLING BEHIND CATCHING-UP 0,0-2 0 2 4 6 8 10 12 Average annual growth rate real R&D expenditures (%) * R&D intensities are for 2005 or most recent year. Trends are calculated as the average yearly growth of real R&D expenditures over a 5- year period. The dotted lines give the mean scores for the group of benchmark countries. The average R&D intensity of each country in 1999-2001 is represented by a square, the 2005 R&D intensity by a circle. Data sources: OECD MSTI database. Data treatments: UNU-MERIT. * R&D intensities are for 2005 or most recent year. Trends are calculated as the average yearly growth of real R&D expenditures over a 5- year period. The dotted lines give the mean scores for the group of benchmark countries. The average R&D intensity of each country in 1999-2001 is represented by a square, the 2005 R&D intensity by a circle. Data sources: OECD MSTI database. Data treatments: UNU-MERIT. The business sector accounts for almost 60% of total R&D spending in the Netherlands, or 5.1 billion euros in 2005. 6 The R&D intensity of the Dutch business sector 1.01% in 2005 also lags behind the intensity of most benchmark countries (Figure 3), and is small in comparison with both the EU25 average (1.3% of GDP) and particularly the average of the benchmark countries (1.8% of GDP). Only in Ireland and Norway is the R&D intensity below that of the Netherlands. In real terms, business R&D has hardly grown at 0.1% per annum. Only in Belgium we observe a lower growth rate. Fastest growing countries are once more South-Korea, Australia, Austria and Ireland. While R&D spending has increased markedly in most countries over the last decade, R&D spending in the Netherlands has remained stable. A major part of the gap can be attributed to the industry structure in the Netherlands, which is characterized by relatively large R&D-extensive sectors. Within the business sector, the industry sector dominates with a spending share of 77% followed by the services sector which accounts for 18% of business R&D spending. Spending shares of industry versus 6 Total R&D expenditures by the business sector are referred to as Business Enterprise Expenditure on R&D or BERD. 11

Figure 4. The Netherlands is among the following nations as far as university R&D expenditure is concerned, but growth is also low University R&D intensities (2005) and 5-year R&D growth rates (1999-2005)* Figure 5. The Netherlands is also falling behind slightly in R&D expenditure by research institutes R&D intensity (2005) and R&D growth rates (1999-2005) of research institutes* R&D-intensity (%) 0,9 FOLLOWERS 0,8 0,7 0,6 DNK SWE CHE FIN Average for benchmark countries AUT CAN LEADERS 0,40 0,35 0,30 0,25 R&D-intensity (%) FOLLOWERS JPN FRA GER FIN AUS EU25 NLD Average for benchmark countries US NOR KOR LEADERS Average for benchmark countries 0,5 0,4 0,3 NLD FRA JPN BEL GER EU25 NOR UK US AUS KOR Average for benchmark countries IRE 0,20 0,15 0,10 0,05 DNK CAN CHE SWE UK AUT BEL IRE FALLING BEHIND CATCHING-UP 0,2 0 2 4 6 8 10 12 14 Average annual growth rate real R&D expenditures (%) * R&D intensities are for 2005 or most recent year. Trends are calculated as the average yearly growth of real R&D expenditures over a 5- year period. The dotted lines give the mean scores for the group of benchmark countries. The average R&D intensity of each country in 1999-2001 is represented by a square, the 2005 R&D intensity by a circle. Data sources: OECD MSTI database. Data treatments: UNU-MERIT. FALLING BEHIND CATCHING-UP 0,00-4 -2 0 2 4 6 8 10 Average annual growth rate real R&D expenditures (%) * R&D intensities are for 2005 or most recent year. Trends are calculated as the average yearly growth of real R&D expenditures over a 5- year period. The dotted lines give the mean scores for the group of benchmark countries. The average R&D intensity of each country in 1999-2001 is represented by a square, the 2005 R&D intensity by a circle. Data sources: OECD MSTI database. Data treatments: UNU-MERIT. services and of large firms versus SMEs (Small and Medium sized Enterprises) have not changed in the past 5 years. The R&D intensity of Dutch business is smaller than that in most reference countries as, due to a shift in the 1990s from R&D intensive manufacturing towards more R&D extensive services, the manufacturing sector has a smaller share in the business sector than in most other countries. There has also been a shift within manufacturing from high-tech manufacturing to low-tech manufacturing. 7 The Dutch business R&D intensity is among the lowest, having dropped from 1.10% in 1999. But we also observe a decrease in business R&D intensities in seven other countries: Belgium, Canada, France, Norway, the UK, the US and Sweden. So there appears to be a more general trend of a declining intensity of business R&D. Almost 75% of business R&D is done by large firms with 250 or more employees, the remaining 25% is done by SMEs. The Netherlands is a small open economy and its industry is dominated by a few (very) large multinational companies: Philips, ASML, Akzo Nobel, NXP, Shell, DSM, Océ and Unilever. Collectively, these eight R&D-intensive multinationals spent about 2.7 billion euros on R&D in 2005, which amounts to more than half of total business R&D expenditures in the Netherlands. In a world where R&D activities are globalizing and firms are displacing R&D activities to other countries, we find that Philips, Akzo Nobel and ASML have actually increased their share of global R&D activities carried out in the Netherlands between 2001 and 2005. For the eight multinationals as a group, the share of worldwide R&D performed in the Netherlands has also increased. Only Shell and DSM show declines. 7 Compare the previous NOWT 2005 report for a detailed analysis of both shifts. 12

Figure 6. The Netherlands exhibits an average pattern of R&D expenditures, but government expenditures on university R&D are relatively high Total R&D expenditure by source of funding, 2003* 100 80 60 (%) 19 5 8 23 21 20 14 8 10 10 25 4 5 16 27 7 18 24 8 11 28 11 21 15 9 11 21 10 3 13 13 5 31 7 4 10 13 13 3 18 10 20 4 3 15 8 3 9 17 24 7 5 12 Funds from abroad Other domestic sources General University funds Direct government 40 20 44 45 48 49 49 51 51 57 61 64 65 67 67 69 70 74 75 Private sector 0 UK AUT CAN AUS NOR FRA NLD IRE DNK US SWE BEL GER CHE FIN KOR JPN * Reference year is 2003, except for Switzerland (2000), Belgium (2001), Australia and Austria (2002), Germany, Ireland, and the US (2004). Data sources: OECD MSTI database. Data treatments: UNU-MERIT. The university sector is the second largest sector in terms of R&D spending. 8 In 2005 the 14 universities in the Netherlands spent 2.5 billion euros on R&D, accounting for 28% of total Dutch R&D expenditures. A large share of this sector s R&D spending is in basic research that serves the advancement of public knowledge by exploring more fundamental and societywide problems related to e.g. predicted shortages of natural resources and problems resulting from global warming. The R&D intensity of the Dutch university sector 0.48% in 2005, just below its 1999 level of 0.51% - is above the average intensity of the benchmark countries and also above that of the EU25 (Figure 4). In real terms R&D spending has hardly grown and in all other countries growth of university R&D spending in real terms is higher. Ireland and Canada show the highest growth rates and increase in R&D intensities. In most countries R&D intensities have increased, only in the Netherlands, France and Japan intensities has fallen. Although R&D spending relative to GDP by Dutch universities is still above average, the lagging growth of real R&D spending poses a threat to this position. The research institutes sector accounted for 14% of total Dutch R&D spending or about 1.1 billion euros 9. The largest institutes are TNO (Netherlands Organisation for Applied Scientific Research), NWO (Netherlands Organisation for Scientific Research), KNAW (Royal Netherlands Academy of Arts and Sciences) and the GTIs (Large Technological Institutes). TNO spends most on R&D: 355 million euros in 2005 or almost 30% of R&D spending by all research institutes. The R&D intensity of 0.24% for Dutch research institutes in 2005 is below the average of the benchmark countries and similar to that of the EU25 (Figure 5). France, Germany and South-Korea have the highest R&D intensities, the fastest growth in real spending is observed for Ireland, South Korea and Belgium. The R&D intensity has fallen in Australia, Canada, Denmark, Finland and Japan, has remained more or less the same in the Netherlands, Norway, Austria and Switzerland, has increased somewhat in Germany, France, Ireland, the UK and the US, and has increased significantly in Belgium, South Korea and Sweden. Figure 6 provides a breakdown of the sources of R&D funding for total R&D expenditures. In most countries the business sector has the largest funding share, and in Japan and South- Korea the business sector pays for about 75 cents of every euro spent on R&D. The funding share in the European countries is between 44% for the UK and 70% for Finland. The business funding share in the Netherlands is 51%, which is below the average for the benchmark countries. The government is an important financer of R&D, with a funding share of 35% or higher, in the following countries: Canada, the Netherlands, France, Norway and Australia. Foreign funding is relatively important with a funding share of 10% or more in Austria, Belgium, Denmark, the Netherlands and the UK. When 8 Total R&D expenditures by the university or higher education sector are referred to as Higher Education Expenditure on R&D or HERD. 9 Total R&D expenditures by the research institutes sector are referred to as Government Intramural Expenditure on R&D or GOVERD. 13

Figure 7. Universities in the Netherlands receive a relatively large share of their R&D-funding from government sources University R&D expenditure by source of funding, 2003* (%) 100 80 47 60 3 40 40 27 9 27 8 21 6 37 27 5 68 7 12 13 69 6 18 5 45 15 14 71 8 3 6 45 8 6 3 41 11 5 68 6 8 3 58 13 61 4 7 75 3 5 5 64 3 5 59 3 5 58 4 4 74 Foreign sources Other domestic sources Private sector General University funds Government 20 0 11 36 28 26 38 42 16 26 24 12 23 29 32 17 JPN CAN UK US BEL SWE KOR FIN IRE CHE DNK GER NLD NOR AUS FRA AUT * Reference year is 2003, except for Switzerland (2000), Belgium (2001), Australia and Austria (2002), Germany, Ireland, and the US (2004). In the case of Belgium and Korea, the share of government funding includes General university funds. Source: OECD MSTI database. Data treatments: UNU-MERIT. Figure 8. Public sector research institutes in the Netherlands receive a relatively large share of their R&D-funding from the private sector R&D expenditure of public sector research institutes by source of funding, 2003* 100 80 60 40 (%) 21 12 67 11 5 16 68 10 1 14 76 10 10 80 2 13 5 80 4 3 9 84 9 5 2 85 7 1 6 86 1 4 6 89 2 3 93 2 4 94 5 94 2 2 1 95 4 96 1 99 100 100 Foreign sources Other domestic sources Private sector Government 20 0 BEL NLD FIN NOR AUS UK DNK FRA AUT GER IRE KOR SWE CAN JPN CHE US * Reference year is 2003, except for Switzerland (2000), Belgium (2001), Australia and Austria (2002), Germany, Ireland, and the US (2004). Source: OECD MSTI database. Data treatments: UNU-MERIT. comparing the findings in Figures 2 and 6, it seems that those countries with a lower R&D intensity have relatively small funding shares for the business sector. The relatively large fraction of General university funds within the R&D spending profile of the Netherlands requires a closer look at university R&D. Figure 7 displays the breakdown of university R&D expenditure by source of funding, which indicates that Dutch universities receive three quarters of their funding from this source, which is significantly more than in all other countries. On top of which, an additional 12% of the expenditure is funded by other government sources. The Dutch universities receive 7% of their funds from the private sector, which is a relatively large fraction within this set of benchmark countries. The funding profile is quite different in the case of the public research institutes (Figure 8). Here we find that the share of government support is among the lowest in the Netherlands; only 68% is government funding. Moreover, these research institutes manage to attract 16% of their funding from industry, which puts the Netherlands in the number one rank in this 14

Table 9. Clustering of countries based on their R&D expenditures and funding (2000-2004) Expenditures Business sector (%): Research institutes (%): Universities (%): Cluster A Low (56%) High (14%) High (29%) Cluster B High (70%) Low (7%) High (23%) Cluster C High (71%) High (15%) Low (15%) Cluster 1 Business sector: low (49%) Government: high (39%) Abroad: low (8%) Australia ( /++) Canada ( /+) Netherlands ( / ) Norway ( /+) France (-/-) Cluster 2 Business sector: low (44%) Government: high (32%) Abroad: high (20%) UK ( /-) Austria (0/++) Funding Cluster 3 Business sector: a (63%) Government: high (29%) Abroad: low (6%) Denmark (0/+) Ireland ( /++) Sweden (++/0) Germany US (0/-) Cluster 4 Business sector: high (72%) Government: low (23%) Abroad: low (2%) Finland (++/+) Switzerland (+/+) Japan (++/+) South-Korea (+/++) Cluster 5 Business sector: high (67%) Government: low (18%) Abroad: high (13%) Belgium ( / ) Between brackets for every country, the R&D intensity and growth of real R&D spending are shown relative to the average scores for the benchmark countries: ++: strongly above average. +: above average, 0: average; -: below average, : strongly below average. Data source: OECD. Data treatments: UNU-MERIT. respect thus indicating a relatively strong orientation of these research institutes toward exploitation of knowledge, skills and facilities in the marketplace. The largest of these research institutes is TNO (see box 2 in Chapter 3). How typical is the R&D expenditure profile of the Netherlands? Can it be compared to one of the other benchmark countries? To address this question, countries were classified in different groups based on differences in spending shares (Figure 2) and funding shares (Figure 6). The country groups are shown in Table 9. In the table rows we find the clusters based on the funding shares between business, government and funding coming from abroad. In the columns, we find three groups or clusters based on the spending shares between the business sector, research institutes and universities. In expenditure cluster A are the countries with a relatively small expenditure share for the business sector: Australia, Canada, Netherlands, Norway and the UK. Expenditure cluster B countries are characterised by a relatively high spending share of the business sector and universities. In this cluster we find the smaller European countries. Expenditure cluster C countries are characterised by a relatively high spending share of the business sector and research institutes. With the exception of the UK, we find all large(r) countries in this cluster. The first funding cluster includes those countries with a high government funding share and a small foreign funding share: Australia, Canada, France, Netherlands and Norway. The second cluster includes Austria and the UK: both countries have a high foreign funding share and government funding share. 15

Figure 10. The Netherlands has a relatively large share of knowledge workers Human Resources in Science and Technology (% of population aged 25-64)* Figure 12. There are relatively few female researchers in the Netherlands Share of female researchers as % of all researchers (2003)* 25 20 15 10 5 (%) 10 14 14 14 15 16 18 18 19 21 22 23 (%) 60 50 40 30 27 25 20 20 17 12 10 7 9 34 34 29 30 31 25 20 18 10 53 44 37 38 39 25 20 19 18 0 0 AUT FRA GER IRE UK BEL CHE NLD FIN SWE NOR DNK 2000 2006 * Relates to HRST-Core personnel with a tertiary education and a science & technology occupation. Data source: EUROSTAT. Data treatments: UNU-MERIT. * Data for 2003, except 2002 for Finland, Austria and Switzerland, and 2000 for the UK. No data available for private sector researchers in the UK. Source: She Figures 2006. Data treatments: UNU-MERIT. Figure 11. The share of researchers is quite low in the Netherlands Share of researchers within the total labour force ( ) 20 ( ) base year 15 10 5 4,8 5,8 5,8 6,1 6,8 7,1 7,2 7,3 7,4 7,6 7,9 9,0 9,3 9,8 10,6 11,9 15,0 2005 increase 0-5 NLD IRE CHE UK GER AUT BEL CAN FRA KOR AUS NOR US DNK JPN SWE FIN JPN GER CHE NLD AUT DNK FRA BEL UK NOR IRE SWE FIN Business sector Universities Data source: OECD MSTI database. Data treatments: UNU-MERIT. The baseline refers to the period 1999-2001, except for Denmark and Austria (2002), and France 2000-2001). In the case of Switzerland, Australia and Canada the reference year is 2004 and 1998-2000 as a baseline; for Switzerland the baseline is 2000. Baseline data were unavailable for Finland, Netherlands and Sweden. The third cluster includes those countries with an average funding share of the business sector: Denmark, Germany, Ireland, the US and Sweden. Countries with a high business funding share and a small foreign funding share are included in the fourth cluster: Finland, Japan, South-Korea and Switzerland. Finally, Belgium is a separate case (cluster 5) with high funding share for both business and funding coming from abroad. Table 9 shows that the Netherlands can be best compared to Australia, Canada and Norway. These countries share similar spending and funding patterns. Although the Netherlands is 16

comparable in terms of its R&D intensity with Australia, Canada and Norway, it does have a lower growth of real R&D spending. A similar pattern can be observed for the universities and the business sector. Only for the research institutes, growth in real R&D spending for the Netherlands is above average. Countries with a high R&D intensity and the highest growth in real R&D spending are characterised by both a high spending and funding share of the business sector. Belgium is an exception here with a similar lagging trend performance as the Netherlands. Based on these characteristics and country groups shown, one may conclude that, if the Netherlands wants to raise its R&D spending to an internationally comparable level, the Dutch business sector should increase its spending and funding share in Dutch R&D spending. 2.2 Human resources devoted to R&D A skilled work force is the backbone of any advanced knowledge-intensive economy, a pool of employees and individuals who can apply their advanced knowledge and know-how in their daily line of work. The availability of highly skilled human capital is generally considered to be an important factor in, if not the prime reason for sustaining knowledge-based economic activity, and for multinational enterprises to (re)locate R&D activities in foreign countries like the Netherlands. The knowledge intensity of the Dutch work force is slightly more than average when compared to other Western European nations. The statistics in Figure 10 indicate that these knowledge workers account for 10-25% of the work force in the Western European nations. In the Netherlands, we find 18% of the working-age population (between 25 and 65 years of age) belonging to this group of professionals who work in R&D-related jobs and have at least one tertiary education degree. However, only a minor fraction of the knowledge workers, those designated as R&D personnel, are actually involved with the development, storage, diffusion and immediate applications of new scientific and technological knowledge. These people can be found both within the manufacturing sector, as well as the services sector and notably also in the university sector. The universities often play a key role, both as a supplier and an employer of R&D staff. University graduates and PhD graduates constitute a pool of human resources out of which companies and research institutes can select their future researchers and technical personnel, in particular those who graduated in natural sciences and engineering. Researchers - either academics or staff at research institutes and R&D-intensive technology companies - are often at the core of knowledge creation processes; they are the ones that push back the knowledge frontiers and explore the potential of new information and novel techniques. The Netherlands has a relatively low share of researchers only 4.8 of the labour force, much less than the Western European benchmark countries (Figure 11). Part of this gap is explained by the relatively small size of the research-intensive manufacturing sectors in the Netherlands, especially the small number of sciencebased technology companies that employ researchers for their R&D units. The Netherlands is also below the European average in terms of female research staff both within the university sector and within the private sector (Figure 12). Several structural problems exist with respect to R&D-relevant human resources in the Netherlands. One of them is the labour market perspective for university graduates and researchers, which is currently characterized by a shortage of attractive job opportunities, unattractive working conditions and career prospects alongside a declining interest among Dutch youth to pursue a scientific career. Currently, the influx of young highly-skilled individuals alleviates some of the shortages, although overall more Dutch highly-skilled persons work abroad than foreign ones in the Netherlands. The Netherlands is one of the few OECD countries that does not profit from a net brain gain. Moreover, the labour participation has declined among university graduates since 2000, notwithstanding a recent partial recovery. The unemployment rate of university graduates is, however, among the lowest internationally. As regards the female work force in the Netherlands, labour participation of highly-skilled women is structurally lower than among their male counterparts. The fraction of female researchers is growing gradually, but they are still significantly underrepresented within the Dutch university research staff, especially within the higher ranks. These phenomena, which are occurring in many other western countries, threaten the adequate supply of human resources in the Dutch R&D system, especially in view of the internationalization of labour markets for highly qualified knowledge workers and the ageing of highly skilled personnel. However, at present, the age distribution of scientific personnel at Dutch universities seems well-balanced. Having a sufficient large pool of researchers within a national R&D system either domestically trained or from abroad - is an important framework condition in terms of guaranteeing a critical mass of brain power for generating new knowledge, making scientific discoveries, and producing R&D-based inventions and technological innovations. Thus far, the comparative shortfall of research staff in the Netherlands has not negatively affected overall productivity levels. This relatively small R&D force can boast competitive output levels in terms of research publications and patents - see Figures 13, 14 and 18 in the next chapter that will deal with R&D performance. 17

3R&D performance 3.1 Research publication output The Dutch science base is strong. The international research performance is competitive at the global level, especially with respect to basic (exploratory) longer-term research. The Netherlands accounts for a mere 0.25% of the world population, but produces about 2 per cent of the world s research publication output in international scientific journals. Those journal publications with Dutch author addresses receive about 3 per cent of all citations worldwide. This achievement places the Netherlands among the world s highest ranking countries in terms of international scientific impact, an indicator of international visibility and research quality. The large majority of those publications are research articles reporting the results of original research with a relevance to a wider, often global, scientific community (see Box 1 for more details about the bibliometric approach for measuring scientific performance). The science base in the Netherlands is also very productive. Dutch researchers are major contributors at the international frontiers of basic science. Their total publication output now stands at an annual average of some 25,000 publications in the international journals (Figure 13). This translates to a daily average of almost 70 publications that are published in international peer-refereed scientific and technical journals. The output has increased by some 40% over the last 10-15 years, which is in line with the growth rates of output produced by other European countries within the international research literature. Given the relatively small numbers of researchers in the Netherlands (see Figure 11), the research productivity rates (publication output per head counts) are quite high in the Netherlands compared to most of the benchmark countries (Figure 14). 10 The Netherlands is among the top 3 nations in terms of productivity per researcher. In terms of the productivity of Dutch public sector researchers (academics mainly), the Netherlands is on an equal footing with the US, the UK and Switzerland. 10 Statistic Netherlands is planning to revise the data for the number of researchers in the Netherlands. Following a past change in methodology, research assistants were not included within these numbers since 2002. By reclassifying these research assistants the number of Dutch researchers will be adjusted upwards. This might have implications for the relative ranking of Dutch research productivity as shown in Figure 14. 18

The prominence of Dutch science at the international level is also manifest in the relative number of citations to research articles (co-)authored by researchers based in the Netherlands. These articles receive 35% more citations than the worldwide average. Again, the Netherlands can boast on a top three ranking worldwide in terms of the (field-corrected) citation impact score (Figure 15), surpassed only by the US and Switzerland. Even the best performing countries may have fields of science in which they are less strong, either in terms of research activity and publication output, or in terms of citation impact. A breakdown of a country s research performance by its share of the world s total output per field, in conjunction with the international citation impact of those publications, provides a general overview a finger print of the field-wise performance of its national science base. Table 16 exhibits the finger print of the Dutch science base. Several fields have extremely high citation scores (>40% above world average), yet belowaverage shares within the total national publication output (>10% below world average), thus indicating excellent research performances with relatively modest means. This category of top performers includes Chemistry and chemical engineering, but also fields in the social and behavioural sciences, and arts and humanities. The fields of Clinical medicine and Agricultural and food sciences combine a high citation impact score with a relatively large publication output. At the other end of the spectrum are Law and criminology and History, philosophy, and religion, fields with low impact and output levels belonging to the arts and humanities (see Box 1 on the limited scope for meaningful interpretation of this particular finding). Box 1 Measuring scientific performance: research output and impact Research publications are one of the key indicators available for evaluating and assessing science. Output from research includes trained personnel, advances in knowledge (new products, methods), patents and scientific articles. Publication output counts have traditionally been used as a proxy measure for scientific production and productivity of universities, public research centres, companies, individuals or nations. In most cases these articles refer to exploratory basic research, but a fair share will stem from applications-oriented strategic research or applied re- search. The central purpose of these publications is the presentation or discussion of scientific data, theory, methods, apparatus or experiments. The other research-related publications comprise review articles, research notes and research letters, which include commentaries and overviews of original research. In many scientific fields, articles are also crucial for researchers international visibility and career advancement. The volume of research articles published worldwide is a key indicator since these publications are the main means of disseminating research results and validating their quality through a peer-review process. International comparisons often rely on counts pertaining to research articles within peer-reviewed international scientific journals, in most cases the journals indexed by Thomson Scientific for their Citation Index databases such as the web-based Web of Science database. This set now comprises some 12,000 journals of which most are in English. Moreover, the propensity to publish in these journals differs across countries and across scientific fields, distorting the relationship between publication-based indicators and the underlying scientific activities and achievements. Although these databases provide good international coverage, including electronic journals, they do not take into account journals of regional or local importance. Journals are grouped into fields of science by Thomson Scientific according to similarities in content. However, their coverage of the worldwide research literature is limited in several main fields, such as Mathematics and Computer Science, Engineering Sciences, Social and Behavioural Sciences, and especially in the case of the Arts and Humanities where findings should be interpreted with great caution. Research articles are attributed to countries according to the author s institutional affiliation at the time of publication as listed on the publication. Publications are considered institutionally co-authored only if their authors are from different main institutions. The same logic applies to cross-sectoral or international collaboration. Research is an accumulative activity, building on results and achievements from the past. The footnotes, end notes or reference lists within research articles acknowledge those sources. Research publications with a large impact on follow-up research are usually heavily cited as a source. The citation frequency is an attestation to intellec- 19

Figure 13. The Netherlands produces some 25,000 research publications per year Publication output volumes and trends: 1993-1996 vs. 2003-2006 1.300.000 Publication output % increase 450 1.200.000 1.100.000 400 1.000.000 350 900.000 800.000 700.000 300 250 600.000 500.000 400.000 200 150 300.000 200.000 100.000 0 100 50 0 US UK JPN GER FRA CAN AUS KOR NLD SWE CHE BEL DNK AUT FIN NOR IRE Output 1993-1996 Output 2003-2006 Increase publication output in 93-96 versus 03-06 (%) * Left axis publication output; right axis- % increase publication output in 2003-06 versus 1993-96. Source: CWTS/Thomson Scientific Web of Science database. Data treatments: CWTS. tual influence, and international citations highlight the visibility of scientific research beyond institutional or national boundaries. The relative prominence of research performing entities (individuals, organisations or countries) is reflected in (relative) number of citations in the international research literature to their publication output. As citation frequency levels tend to differ in terms of (sub)fields of science, these frequency counts per publication are normalized by the field-specific average citation frequencies. These field-corrected citation impact scores enable straightforward comparisons between entities irrespective of the fields of science to which they relate. Universities and public sector research institutes clearly have an important role to play in maintaining excellent research capabilities and increasing the competitiveness of the Dutch knowledge-based economy. Some three-quarters of the publication output is produced by scientists and scholars employed at the 14 universities in the Netherlands and their medical centres (Table 17). An additional 12% originates from the public sector research institutes. This category is quite diverse, ranging from KNAW institutes dealing with research documentation to Leading Technology Institutes focusing on pharmaceutical R&D in cooperation with industry (Box 2 provides a brief overview of the research universities and the large variety of research institutes). The private sector accounts for about 5% of the total publication output, in large part produced by the research-intensive technology companies in the Netherlands with large R&D units: Philips, Unilever, DSM, Shell and Akzo Nobel Organon. Several large general hospitals also conduct (collaborative) research, often in clinical medicine: they account for 5% of the output. The remaining 1% of the research publication output is produced by governmental organizations. The citation impact of the research within the various institutional sectors of the domestic science base varies from 1.51 within the specialized research institutes (i.e. 51% above world average) down to 1.11 for the small share of research publications produced by the government organizations. The bulk of the research output produced by the university researchers, reaches a level of 1.32. This is almost identical to the impact of the publications (co-)produced by the private sector, mostly R&D staff employed by the large research-intensive multinational companies in the Netherlands. In conclusion, the strength of the science base is not concentrated in a single institutional sector but is quite strong across the board. In short, the public research sector is diverse, comprising a wide variety of research organisations active across the en- 20

Figure 14. Productivity of Dutch researchers is one of the highest worldwide Average research publication output per head count in 2004*,**,*** Relative number of publications 2,5 Output per capita public sector researchers 2,0 1,5 Output per capita public and private researchers Output per 1000 capita total population 1,0 0,5 0,0 KOR JPN FRA GER US AUT IRE BEL UK CAN AUS NLD NOR FIN DNK SWE CHE * The number of researchers is expressed in full-time equivalents. ** Total number of researchers in the public and private sector in 2004 (UK, 1998; Canada and the US, 2002; Netherlands and Sweden, 2003). *** Total number of researchers in the public sector (higher education sector and government sector) in 2004, except: UK, 1998 (data for HE); US, 1999 (HE) and 2002 (government); Canada 2002 (HE and government); Netherlands, 2003 (HE) and Sweden, 2003 (HE and government). Sources: CWTS/Thomson Scientific Web of Science database; OECD MSTI database. Data treatments: CWTS. Figure 15. The Netherlands is among the world s most highly cited nations Citation impact scores and trends: 1993-1996 vs. 2003-2006*,** 1.60 Impact scores % increase 35 Impact 1993-1996 1.40 1.20 1.00 0.80 0.60 30 25 20 15 10 5 0 5 Impact 2003-2006 Change in citation impact score between 93-96 en 03-06 (%) 0.40 CHE US NLD DNK UK CAN NOR BEL SWE IRE FIN GER AUT AUS FRA JPN KOR 10 * Left axis field-normalized citation impact score; right axis- % change in citation impact score between 1993-1996 and 2003-2006. ** The field-normalised citation impact scores are defined as the quantity of citations received by research publications relative to the worldwide average citation impact per field (world average score=1.0). Data relate to publications from 2003-2006, cited during the years 2003-2006. Source: CWTS/Thomson Scientific Web of Science database. Data treatments: CWTS. 21

Table 16. Fingerprint of Dutch science in terms of research fields International research specialisation profile and international citation impact (2003-2006)*,** Citation impact (CI score*) Share of the Netherlands in worldwide publication output (OSI score**) Below average (OSI 0.9) Average (0.9 < OSI <1.1) Above average (OSI 1.1) Very high (CI 1.4) Chemistry and chemical engineering Physics and material science Information and communication science Arts, culture and music Literature High (1.2<CI<1.4) Electronics and Electrical Engineering Civil engineering Instruments and instrumentation Earth science and technology Computer science Environmental science and technology Agricultural and food science Clinical medicine Above average (1.1<CI 1.2) Political science and Public administration Mathematics Mechanical engineering and aerospace Energy and fuels General and industrial engineering Biological science Fundamental life science Educational science Biomedical science Astronomy and Astrophysics Average (0.9<CI 1.1) Sociology and anthropology Basic and experimental medicine Health sciences Language and linguistics Statistics Economics and business Management and planning Psychology Social and behavioural science - interdisciplinary Below average (CI 0.9) Law and criminology History, philosophy, and religion * CI score (2003-2006): field-normalised citation impact scores in 2003-2006, i.e. the quantity of citations to publications in a field of science relative to the worldwide average citation impact per field. World average = 1.0. Relates to publications from 2003-2006, cited during the years 2003-2006. ** OSI score (2003-2006): % share of research publication output per field of the total Dutch publication output, relative to the average share of the same field in the publication output of all benchmark countries (non weighted average of countries). Source: CWTS/Thomson Scientific Web of Science database. Data treatments: CWTS. 22

Box 2 Research universities and research institutes in the Netherlands In the Netherlands there are two main types of regular higher education, namely research universities and universities of applied sciences. There are 14 government-approved research universities in the Netherlands, 13 of which focus on research as well as education (offering bachelors, masters, and PhD degrees). The universities vary in size, with student enrolments ranging from 6,000 to 30,000. The 13 research universities are, in descending order of publication output in 2006: Utrecht University, University of Amsterdam, VU University Amsterdam, Erasmus University Rotterdam, Leiden University, University of Groningen, Radboud University Nijmegen, Wageningen University and Research Centre, Maastricht University, Delft University of Technology, Eindhoven University of Technology, University of Twente and Tilburg University. The universities employed some 18,000 research staff in 2005. The Netherlands is also home of many non-university research institutes. Most of them are publicly funded, such as the institutes funded by the Netherland Organisation for Scientific Research (NWO), which also acts as the national research council across all fields of science. The NWO research institutes are: Institute for Astronomical Research in the Netherlands; The National Research Institute for Mathematics and Computer Science; FOM-Institute for Atomic and Molecular Physics; FOM-Institute for Plasma Physics Rijnhuizen ; National Institute for Nuclear Physics and High Energy Physics; Institute for Dutch History; Royal Netherlands Institute for Sea Research; Netherlands Institute for the Study of Crime and Law Enforcement; and SRON Netherlands Institute for Space Research. These institutes employed a total of approximately 2,100 staff in 2006. The research institutes at the Royal Netherlands Academy of Arts and Sciences (KNAW) carry out basic and strategic research in the life sciences, humanities and social sciences. Some of the institutes also have a scientific service function by forming and managing biological and documentary collections, providing information services and creating other facilities for research. The institute comprises of: Data Archiving & Networked Services; Netherlands Institute for War Documentation; Fryske Akademy; Netherlands Institute of Ecology; Fungal Biodiversity Centre; Netherlands Institute for Neurosciences; Hubrecht Institute; Netherlands Interdisciplinary Demographic Institute; Huygens Institute; Rathenau Institute; International Institute for Social History; Roosevelt Study Center; Interuniversity Cardiology Institute of the Netherlands; Royal Institute of Linguistics and Anthropology; Meertens Institute; Virtual Knowledge Studio for the Humanities and Social Sciences; and Netherlands Institute for Advanced Study in the Humanities and Social Sciences. The KNAW employs a total of approximately 1,100 staff. Other public research institutes in the Netherlands are: TNO, National Institute for Public Health and Environment, Netherlands Cancer Institute, KNMI (weather and climate research institute), and Sanquin (blood related research and services). In addition, there are five Large Technological Institutes and eight Leading Technology Institutes focusing on R&D activities, with links to the private sector and partially funded by industry. Table 17. Research performance at institutional sector level Share of total publication output and international citation impact (2003-2006)*,** % publication output Citation impact score* Universities (including 76% 1.32 medical centres) Public research institutes 12% 1.51 Private sector 5% 1.33 General hospitals 5% 1.22 Government 1% 1.11 * CI score (2003-2006): field-normalised citation impact scores in 2003-2006, i.e. the quantity of citations to publications in a field of science relative to the worldwide average citation impact per field (world average=1.0). Relates to publications from 2003-2006 and citations during the years 2003-2006. ** Excludes international organizations based in the Netherlands (e.g. ESA-ESTEC). Source: CWTS/Thomson Scientific Web of Science database. Data treatments: CWTS. tire range of scientific disciplines, but the large majority of the research publication output is concentrated in 13 universities and a few large research institutes (see Box 2). Although each of these universities and institutes differ in size and scope, they all perform quite well in terms of producing large numbers of publications that generate above average citation impacts on the global scientific community (see Table 18). 23

Table 18. Research performance at the organisational level Publication output and international citation impact scores of the universities and major research institutes (2003-2006)*,** Publication output Citation impact score* Utrecht University 13,588 1.43 University of Amsterdam 11,470 1.40 VU University Amsterdam 9,128 1.39 Erasmus University Rotterdam 8,903 1.56 Leiden University 9,405 1.23 University of Groningen 8,624 1.23 Radboud University Nijmegen 8,558 1.20 Wageningen University and 6,307 1.23 Research Centre Maastricht University 5,761 1.22 Delft University of Technology 5,644 1.40 Eindhoven University of 4,281 1.50 Technology University of Twente 3,116 1.35 Tilburg University 1,151 1.12 TNO 1,962 1.16 National Institute for Public 1,393 1.52 Health and Environment Netherlands Cancer Institute 1,456 1.95 National Research Institute 677 1.25 for Math. and Computer Sci. Netherlands Institute for 513 1.51 Sea Research Institute for Atomic and 578 2.56 Molecular Physics Netherlands Institute of 614 1.35 Ecology * CI score (2003-2006): field-normalised citation impact scores in 2003-2006, i.e. the quantity of citations to publications in a field of science relative to the worldwide average citation impact per field (world average=1.0). Relates to publications from 2003-2006 and citations during the years 2003-2006. ** Main organisations with 500 or more research publications in the period 2003-2006. Source: CWTS/Thomson Scientific Web of Science database. Data treatments: CWTS. Several Dutch universities are high on the worldwide university rankings. 10 A few universities (Rotterdam and Eindhoven) and several research institutes, such as the Institute for Atomic and Molecular Physics and the Netherlands Cancer Institute are cited more than 50% above worldwide average, which puts them amongst the most highly cited worldwide. 3.2 Patent output A strong science base is a necessary condition for an internationally competitive and vibrant national innovation system. But a strong science base is not a sufficient condition for economic success, especially when science-based knowledge, skills and techniques are not (adequately) transferred and exploited for inventions and further applications of commercial value in the (inter)national market place. 11 Box 3 Measuring technological performance: patents There is a positive association between patenting and R&D investment, especially within the high tech industrial sectors. High R&D-intensive industries, such as pharmaceuticals or medical, precision and optical instruments, are among those that patent the most. As a measure of R&D outputs, patenting by industry provides empirical information on industries technological strengths and innovation potential. Patents taken out on technical inventions can be used as a proxy measure of technical advances and technological innovation, especially in technical areas and industrial sectors where patents are used by business companies to protect intellectual property. The importance of patents for protecting knowledge does not depend solely on the level of R&D investments. Differences among industries in terms of the risk of imitation and the extent to which patents enhance competitive advantages in markets (e.g. through technology exchanges and alliances) also affect the use of patents by companies. Patents are attributed to countries according to the geographical location of the applicant (usually a company) or the residence of the inventor(s). Patents applied for by multinational companies may originate from foreign R&D and list inventors from different countries. The statistics shown in this report refer to patents assigned according to the country of inventor(s). 11The Shanghai ranking (http://www.arwu.org/rank/2007/ranking2007.htm), Times Higher Education Supplement ranking (http://www.thes.co.uk/worldrankings/), CWTS/Leiden Ranking (http://www.cwts.nl/cwts/about_ranking.html), and the CHE ranking (www.che-ranking.de). 12 NOWT s Science & Technology Indicators Report specifically excludes innovation indicators and descriptions of the state of affairs in the national innovation system in the Netherlands - for this we refer the reader to indicators reports produced by the Statistics Netherlands, in particular: Kennis en economie 2007, Voorburg/Heerlen, Centraal Bureau voor de Statistiek, 2007. 24

Figure 19. The Netherlands is among the leading nations in international patents Patent output per million inhabitants in 2003 (EPO patent applications; USPTO patents; triad patent applications) CHE GER FIN SWE NLD DNK JPN AUT FRA BEL US UK NOR CAN IRE AUS 0 50 100 150 200 250 300 350 400 450 Relative number of patents EPO USPTO Triad Source: OECD MSTI database. Data treatments: CWTS. The volume of patented inventions science-based, researchrelated or otherwise - produced by the Netherlands sheds some light on the levels of technical inventiveness of the Dutch R&D system, especially within the R&D-intensive manufacturing industries (Box 3 provides a brief explanation about patents as an R&D output measure). Based on the general assumption that the patented inventions and innovations (i.e. commercialized inventions) reflect (causal) relationships between inventiveness and innovation, the patent productivity statistics in Figure 19 also offer a crude reflection of Dutch capacity to produce high-quality technological innovations. The international comparative statistics indicate that the Netherlands is among the top-tier countries that apply for international patents, either through the European Patent Office (EPO), the US Patent and Trademarks Office (USPTO), or collectively at EPO, USPTO and the Japanese Patent Office (the triad patents ). Patenting is particularly important in technologybased industrial sectors like electronics, biotechnology, and pharmaceuticals. The relatively large output of patents from ap- plicants in the Netherlands originates from Philips, with its corporate headquarters and large R&D laboratories in the Netherlands. Philips is one of the most prolific companies in terms of patenting, both in the EPO system as well as the USPTO system. Public sector patenting is gradually increasing in the Netherlands, but the growth is entirely due to the research institutes (Figure 20). The patent output by universities has levelled off after a peak in 2000. The share of EPO patent applications from both the university sector and the research institutes amounts to 6% of the total quantity of patent applications from the Netherlands. This share represents a threefold increase compared to 10 years ago, thus indicating that public sector research organisations have become more oriented toward commercial exploitation of their intellectual property. The remainder of the patent applicants are filed by private sector organisations and business companies, with Philips as the most prolific company. 25

Figure 20. Public sector patenting is gradually increasing in the Netherlands EPO patent applications by universities and public research institutes in the Netherlands Number of Patent applications 160 140 120 100 80 60 40 20 % Patent applications 3,5 3 2,5 2 1,5 1 0,5 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 0 Patent applications by universities* Patent applications by research institutes** Share of universities in total patent output of the Netherlands (%) Share of research institutes in total patent output of the Netherlands (%) Source: Steunpunt O&O Indicatoren, Leuven University. Data treatments: CWTS. * Excluding patent applications of companies affiliated to universities. ** Including patent applications by NWO, the KNAW and Large Technological Institutes. Note: Left axis number of EPO patent applications; right axis - % share of EPO patent applications in the total number of EPO patent applications by public sector and private sector applicants in the Netherlands. 26

4R&D co-operation 4.1 Co-operation between industry and the public sector The previous section has shown that the Netherlands has created a strong science base with a wide range of research-intensive universities, research institutes, medical centres, and business enterprises. R&D co-operation is of pivotal importance given the relatively small size of the science base. Scientific co-operation has always been one of the pillars of the tightly knit research community in the Netherlands. Being at the international frontier of science in many fields also means that the Dutch have a fair share of collaborative efforts with foreign partners. Public-private R&D co-operation is another major driving force, and binding factor, within the Dutch research and innovation system. Cooperation between knowledge creating institutions and firms is vital for applying new knowledge and technologies in innovative products and processes. These networks of connections between the public sector and the private sector, and the intermediate organizations supporting those linkages, encourage the utilization of scientific and technical knowledge produced by universities and other publicly funded research institutes. Befitting this small and densely populated country, the Netherlands benefits from a dense infrastructure of R&D programs - both formal and informal - in which the public sector and business sector participate, and where colleagues (and competitors) meet and co-operate. Box 4 introduces the key institutions in the Netherlands engaged in the transfer of public sector R&D results to the business sector in the Netherlands and elsewhere. Box 4 R&D knowledge transfer institutions The three major actors that are actively involved in transferring research-based knowledge to the private sector are: The Netherlands Organisation for Applied Research (TNO); the Large Technological Institutes (GTI, Grote Technologische Instituten); and the Leading Technology Institutes (TTI, Technologische Top Instituten). TNO is partially funded by the Dutch government. The objective of TNO is to create and translate scientific knowledge into applied knowledge that is useful for the private sector and government agencies. TNO comprises five core areas that focus on quality of life, defence, security and 28

sight in how many innovating companies are engaged in cooperative relationships with universities and research institutes. Figure 21 indicates that these public-private linkages are not particularly frequent in the Netherlands; the Dutch are slightly below EU-average in terms of the share of innovative firms acknowledging collaborative arrangements with public sector research organizations. This outcome reflects the fact that only a relatively small share of innovating companies in the Netherlands are R&D-intensive or sciencepublic safety, science and industry, built environment and geosciences, and ICT. TNO has established, together with universities, some 30 knowledge centres to develop knowledge in fields of relevance to innovating companies. The five GTIs are: ECN, GeoDelft, MARIN, NLR, and WL Delft Hydraulics. They are large specialised institutes for demand-driven R&D in the domains of energy, geodesy, marine sciences, aerospace, and hydraulics, respectively. Their main mission is to transform basic knowledge into applications for both industry and government (the GTIs are partially funded by the Dutch government). The TTIs are network-based organizations specifically aimed at fostering business innovation by improving the innovative capacity and competitive strength of industry in a number of selected fields. They perform (joint) research and offer training of scientists and engineers. The TTIs are co-funded by government and industry. Institutional partnerships between industry and the public research infrastructure are paired with industry relevant fundamental and strategic research of an excellent international standard. The TTIs in operation are: Telematica Instituut; Dutch Polymer Institute; Netherlands Institute for Metals Research; Wageningen Centre for Food Sciences; Top Instituut Pharma; Center for Translational Molecular Medicine; TTI Groene Genetica; and Technologisch Top Instituut Watertechnologie. Dutch public and private organizations collaborate often in joint R&D projects, programs and networks. In doing so, universities and public research organizations contribute albeit often indirectly to innovation processes. The public sector supplies human resources (e.g. recently graduated engineers or PhD graduates), offers access to technical facilities, and provides knowledge and know-how through contract research or consultancy arrangements. Government measures have been taken to increase the volume, intensity and effectiveness of R&D linkages between public research organisations and innovating firms. In addition, policies are developed and implemented to increase the current and prospective supply of scientists and engineers with the goal to make the Netherlands a more attractive location for R&D investments. Figure 21. The Netherlands is less competitive than other European nations in terms of its level of public-private R&D cooperation Share of innovative companies cooperating with universities or public research institutes (2002-2004) 80 70 60 50 40 30 20 10 0 70 60 50 40 30 20 10 0 % % 26 FRA IRE 15 16 SWE 31 31 32 33 DNK NLD DNK 18 18 IRE FRA UK NLD BEL Source: Eurostat CIS4 Survey. Data treatments: UNU-MERIT. 37 average The latest Community Innovation Survey, presenting internationally comparative data on innovation performance and behaviour for EU countries for the years 2002-2004, provides in- 38 average UK 41 SWE BEL 45 NOR 24 24 25 26 26 GER 53 GER 30 AUT 58 AUT 49 NOR 75 FIN 59 FIN 29

Figure 22. The science base of the Netherlands shows a relatively strong linkage with industrial research Share of public-private research publications in the total national publication output (2005-2006)* 12 11 10 9 8 7 6 5 4 3 2 1 0 % CHE JPN DNK SWE AUT NLD BEL FIN KOR GER US CAN UK NOR FRA IRE AUS * Public-private co-publications are defined as publications with at least one author address referring to a university in the Netherlands or a public sector research institute, and at least one address referring to a private sector organization (business company or otherwise) either in the Netherlands or elsewhere. Source: CWTS/Thomson Scientific Web of Science database. Data treatments: CWTS. Figure 23. The US, Germany and the UK are major research co-operation partners Volume of internationally co-authored research publications and trends 1993-1996 vs. 2003-2006* 14.000 13.000 12.000 11.000 10.000 9.000 8.000 7.000 6.000 5.000 4.000 3.000 2.000 1.000 0 Number of publications % 250 1993-1996 2003-2006 200 150 100 50 0 increase US GER UK FRA BEL CHE CAN SWE AUS DNK JPN FIN NOR AUT IRE KOR * Left axis output of joint co-publications; right axis - % increase in international co-publication output between 1993-1996 and 2003-2006. Source: CWTS/Thomson Scientific Web of Science database. Data treatments: CWTS. 30

based. Dutch companies are more likely to team up with research institutes as compared to universities, which is understandable, considering that the research institutes tend to focus on applied and strategic research (see Box 4) of more immediate value to companies. with research partners in Australia, South Korea and Ireland. Relatively high growth rates are also found in the case of Norway, Austria and Finland, again emphasizing the importance of the European dimension in the on-going internationalisation of Dutch science. 4.2 Research co-operation The subset of innovating companies in the Netherlands that are research intensive tend to co-operate with (local) universities and research institutes. Part of their joint research efforts especially those dealing with fundamental research and clinical research produce results that are published in the international research literature. The rate of occurrence of these jointly authored research publications is indicative for the magnitude and intensity of public-private research cooperation (notably in the science-dependent industrial sectors, such as biotechnology, pharmaceuticals, chemicals and electronics). Figure 22 shows the relative share of these public-private co-publications within the total publication output of the Netherlands, alongside the shares in the benchmark countries. The Netherlands is among the higher ranking countries with an 8.5% share, on par with Sweden and Austria. At the top are Switzerland, Japan and Denmark with shares around 10%. Note that the partnering companies are not necessarily local; particularly in the case of the smaller countries like Switzerland, Denmark and the Netherlands with a relatively small number of potential partners, one may expect to find foreign partners as well. The international dimension of research cooperation as a whole is reflected in Figure 23, which shows the preferred partner countries of Dutch researchers. The US, the UK and Germany are at the top. In all three cases, the number of copublications with researchers in these countries has increased by 100% or more during the last 15 years. There were more than 13,000 US-Dutch co-publications during the years 2003-2006 with at least one author address in the US and one in the Netherlands, up from 6,000 in 1993-1996. These findings reflect structural features of global science: the US is the largest knowledge producer and contributor of research publications worldwide (and therefore more likely to become a partner of Dutch researchers) and international scientific cooperation has increased significantly worldwide (thus producing a large increase of international co-publications in many countries). Note that the growth rates with the European partners are much higher than the US-Dutch ties - a sign of integration processes within the European science base. The interesting features are deviations from expected patterns, such as the relatively strong ties with France and Belgium, and the extremely large growth rates from a low base line 31

5General conclusio Like many advanced economies, the economy of the Netherlands is becoming increasingly knowledge based ; a rising share of the products and services for local, regional and global markets are based on improvements and innovations derived from the application of new knowledge or advanced skills. The Netherlands has one of the most competitive economies worldwide, ranked 10 th in the Global Competitiveness Index. Both the educational system and the R&D system contribute to the knowledge intensity of Dutch society and the economy. The Netherlands spent about 8.8 billion euros on R&D in 2005. Its relative level of spending, measured by its R&D intensity, is 1.73% of GDP. However, the Dutch R&D intensity is below average compared to the benchmark countries: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Japan, Korea, the Netherlands, Norway, Sweden, Switzerland, United Kingdom, and the United States. The Dutch growth rate of real R&D expenditures is much lower and it is falling behind. The business sector accounts for almost 60% of total R&D spending in the Netherlands. The R&D intensity of the Dutch business sector 1.01% in 2005 also lags behind the intensity of most benchmark countries. A major part of this gap can be attributed to the industry structure, which is characterized by relatively large R&D-extensive sectors in the Netherlands. The knowledge workers account for 10-25% of the work force in the Western European nations. The knowledge intensity of the Dutch work force is slightly more than average, with 18% of the working-age population between 25 and 65 years of age belonging to this group of professionals who work in R&D-related jobs and have at least one tertiary education degree. However, the share of researchers only 4.8 of the labour force, is much less than the Western European benchmark countries. Dutch researchers are very productive: the Netherlands accounts for a mere 0.25% of the world population, but produces about two per cent of the world s research publication output in international scientific journals. The Netherlands is among the top three nations in terms of productivity per researcher. Those publications receive about three per cent of all citations worldwide. The Netherlands can boast on a top three ranking worldwide in terms of citation impact, receiving 32

ns 35% more citations than the worldwide average. Several of the universities and research institutes are amongst the best worldwide. The Netherlands has created a strong science base with a wide range of research-intensive universities, research institutes, medical centres, and business enterprises. R&D co-operation is of pivotal importance given the relatively small size of the science base. Befitting this small and densely populated country, the Netherlands benefits from a dense infrastructure of R&D programs - both formal and informal - in which the public sector and business sector participate, and where colleagues (and competitors) meet and co-operate. Owing to the fact that only a relatively small share of innovating Dutch companies are R&D-intensive or science-based, the Netherlands is slightly below EU-average in terms of the share of innovative Dutch firms acknowledging collaborative arrangements with public sector research organizations. Focusing on the subset of innovating companies in the Netherlands that are research intensive and tend to co-operate with (local) universities and research institutes, we find the Netherlands among the higher ranking countries. The international comparative statistics indicate that the Netherlands is among the top-tier countries that apply for international patents. Philips, the electronics company, is one of the most prolific companies in terms of patenting. Public sector patenting has gradually increased in the Netherlands during the last 10 years. Summarizing, the Netherlands is amongst the world s leading nations in terms of science and technology with a comparatively high productivity levels based on modest levels of R&D investment, which makes the R&D system a relatively efficient one. It is an open question whether or not the Netherlands will be able to sustain this level of performance, especially in view of the fact that the Netherlands R&D system is affected by (projected) shortages in human resources and seems to be loosing ground in terms of investments in R&D and innovation that are not keeping pace with the rapid growth rates in several other advanced nations. 33

List of abbreviations AUS AUT BEL CAN CHE CIS4 DNK EC EPO EU FIN FRA GDP GER GTIs IRE JPN KNAW KOR MSTI NLD NOR NOWT NWO OECD R&D SME SWE TNO UK US USPTO Australia Austria Belgium Canada Switzerland Fourth Community Innovation Survey Denmark European Commission European Patent Office European Union Finland France Gross Domestic Product (Gross National Product) Germany Large Technological Institutes Ireland Japan Royal Netherlands Academy of Arts and Sciences Republic of Korea (South Korea) Main Science and Technology Indicators (OECD database) Netherlands Norway Nederlands Observatorium van Wetenschap en Technologie (Netherlands Observatory of Science and Technology) Netherlands Organisation for Scientific Research Organisation for Economic Co-operation and Development Research and Development Small or medium sized enterprise Sweden Netherlands Organisation for Applied Scientific Research United Kingdom United States of America US Patent and Trademarks Office 34

Acknowledgements This summary is based on the Dutch-language full report Wetenschaps- en Technologie- Indicatoren 2008 (Science and Technology Indicators 2008) that was released in March 2008. The work on this report has benefited from many valuable inputs from the members of NOWT s advisory board: Cornelis van Bochove (Ministry of Education, Culture and Science, chairman of the board). Jan van Steen (Ministry of Education, Culture and Science, secretary of the board), Luuk Klomp (Ministry of Economic Affairs), Bert Minne (CPB), Joke van den Bandt-Stel (VNO-NCW), Peter Baggen (VSNU), Mariken Elsen, Rien de Jonge en Margreet Bouma (NWO), Ans Vollering (KNAW), Jan Vogel (TNO), and Peter van den Besselaar (Rathenau Institute). Other significant contributions to data collection and data analysis were made by Peter van den Berg (SenterNovem), Clara Calero (CWTS), Bart van Looy (Steunpunt O&O Indicatoren, Cath. Univ. Leuven), Rik Timens (Ministry of Economic Affairs), Arjan Wolters (SenterNovem) and Herman Pijpers (UNU-MERIT). We would also like to express our gratitude to Minna Kanerva (UNU-MERIT) for editorial work on this summary. 35

36

Ministry of Education, Culture and Science in the Netherlands Phone: +31-70.412.34.56 Edition: March 2008 Production: Information Department, Leo Wijnhoven/Jan van Steen Design: Wim Zaat, Moerkapelle Printed by: Koninklijke De Swart, The Hague See also: www.minocw.nl/wetenschap 38.034/550