headline indicators for innovation

Transcription

1 Elements for the setting-up of headline indicators for innovation in support of the Europe 2020 strategy Report of the High Level Panel on the measurement of Innovation Research and Innovation

2 EUROPEAN COMMISSION Directorate-General for Research and Innovation Directorate RTD Directorate for Research and Innovation Unit C.6 Economic Analysis and Indicators Contact: Pierre Vigier European Commission B-1049 Brussels

3 EUROPEAN COMMISSION Elements for the setting-up of headline indicators for innovation in support of the Europe 2020 strategy Report of the High Level Panel on the measurement of Innovation Brussels, September 30, Directorate-General for Research and Innovation

4 EUROPE DIRECT is a service to help you find answers to your questions about the European Union Freephone number (*): (*) Certain mobile telephone operators do not allow access to numbers or these calls may be billed LEGAL NOTICE Neither the European Commission nor any person acting on behalf of the Commission is responsible for the use which might be made of the following information. The views expressed in this publication are the sole responsibility of the author and do not necessarily reflect the views of the European Commission. More information on the European Union is available on the Internet ( Cataloguing data can be found at the end of this publication. Luxembourg: Publications Office of the European Union, 2013 ISSN ISBN doi /15087 European Union, 2013 Reproduction is authorised provided the source is acknowledged. Cover Images: earth, # , Source: Shutterstock.com; bottom globe, PaulPaladin # , Source: Fotolia.com

5 Contents I. INTRODUCTION 4 II. DESIRABLE PROPERTIES OF INDICATORS 5 III. INDICATORS IDENTIFIED 6 IV. BENCHMARKING EUROPE AND THE MEMBER STATES 10 V. OPERATIONAL SCENARIOS 22 VI. FORWARD 24 TECHNICAL ANNEX TO SECTION III 28 COMPOSITION OF THE PANEL 30 3

6 Report of the High Level Panel on the measurement of Innovation established by Ms Máire Geoghegan-Quinn, European Commissioner for Research and Innovation I. INTRODUCTION With the Europe 2020 strategy, the European Union is reasserting its commitment to the goal of a dynamic, sustainable, knowledge-based ( smarter ) economy. At a time when short-term demands arising from the management of the economic crisis could easily dominate the landscape of policy initiatives this is to be commended and applauded. It is now more important than ever to maintain a clear focus on the horizon. Ever since J.A. Schumpeter made the argument with much conviction, it has been accepted that innovation is a basic ingredient of a dynamic economy. Both experience and research have made it abundantly clear that the objectives of the Europe 2020 strategy will not be achieved if Europe does not become a hotbed of innovation, not only in its market activities but also in its institutions and public administrations. The Europe 2020 strategy strongly reflects this view and is correspondingly ambitious. Sound policy-making, including benchmarking and the setting of targets, requires that the state of innovation be adequately measured. This is not a simple matter as innovation is a phenomenon with many facets. In the Lisbon agenda R&D intensity (investment in R&D as a % of GDP) served as a proxy. There is correlation, and causation, between the state of innovation and this well tested and widely available input measure of the state of knowledge generation 1, and for this reason we believe that the decision to retain the 3% R&D intensity target in the Europe 2020 strategy is entirely correct. Yet, there is also merit in the raison d'être of this High Level Panel (HLP): the awareness that the input R&D intensity indicator needs to be complemented by some additional output-oriented indicators. A complete picture should indeed have both. The remit of this HLP is therefore to formulate some headline output-oriented innovation indicator(s) which is (are) adequate, in the context of the Europe 2020 strategy, for policy guidance purposes. This is the objective of this Report 2. A significant proviso is, however, called for: ours is not a straightforward task. On the one hand, we face the conceptual difficulties, still largely unresolved, arising from the need to put precise demarcation lines on what constitutes innovative activity. On the other hand, there are extremely severe limitations in the availability of data. In consequence, at this stage our proposals will have to be imperfect. But we hope they are still useful. 1 See, for example, J. Mairesse and P. Mohnen: Using Innovation Surveys for Econometric Analysis, in Hall, B.H. and N. Rosenberg (2010) Handbook of the Economics of Innovation, Elsevier. 2 Not aiming at giving an exhaustive picture of the innovation phenomenon such indicator(s) will of course need to be completed by detailed scoreboards and analysis. 4

7 The Report is structured as follows. Section II lists a variety of desirable properties that proposed indicators should ideally satisfy. Section III lays out in a non-technical manner a number of pragmatic proposals 3. Section IV moves to a comparative presentation and benchmarking of the different indicators proposed. Section V discusses two operational scenarios for the pursuance of the remit of the HLG. Section VI is very important: it lists, describes and discusses some of the research and data collection programs that would need to be carried out in the future, and in a continuous manner. II. DESIRABLE PROPERTIES OF INDICATORS An indicator, or a set of indicators, intended to measure progress and possibly to serve as the basis for the establishment of challenging European targets for 2020 has to satisfy, ideally, a number of desirable properties. We put here on record those properties which we believe are essential, in the sense that without them it would be very difficult to formulate objectives that have a good chance of enduring, i.e. to serve as policy guides and to inspire the citizenry of Europe. We say essential rather than indispensable because there is an element of a wish-list in what follows. In practice we will need to make compromises. Satisfying each and every one of the desirable properties would be very difficult. The properties are: 1. Simple and understandable. An indicator, perhaps coupled with a corresponding target, has to be easily understandable, so that the intended audience can intuitively relate to it. It can have in fact, it is desirable that it has a precise technical meaning but it should not be a technical concept in itself. As for the targets, if at all possible, they should have the sharpness and mobilizing power of a good call for action. They should be transparent in its meaning and be capable of stirring the imagination of the European public. The 20/20/20 target in the fight against man-made global warming constitutes an excellent example. This criterion led the HLP to exclude a composite indicator that would have been configured from a number of interdependent, but possibly disparate, sub-indicators. 2. Sizable and direct. The indicator should cover a significant share of the issue at hand. It should be sizable and relevant to its substance while, at the same time, rich enough to indirectly implicate some of the framework conditions that potentially sustain an innovative society. 3. Objective. Indicators based on qualitative answers to survey questionnaires are very important and constitute an irreplaceable tool for the in-depth analysis of innovation issues. Yet, for the purpose at hand, where the sturdiness of the indicators has much to be recommended, we believe that the basic indicators have to be built on hard data as much as possible. 4. Presently computable. The formulation of an indicator could serve as a stimulus for the gathering and processing of the necessary data to produce it, which could be initially unavailable. This may be the case in connection with some of our proposals, for instance in order to extend the geographical coverage of existing data. Ideally, this should be the exception rather than the rule. In principle, each indicator selected for the launching phase of the Europe 2020 strategy should be available relatively quickly, be easily measurable and based on available data. Of course, relatively quickly is a 3 A technical annex provides detailed definitions 5

8 relative term. In some instances we shall view two years as being under the purview of this clause. 5. Stable. By this term we mean two things. Firstly, the indicators should be sufficiently recognized and tested that they can last for ten years without modification. It would not make much sense to have to confront a redefinition of the indicators in mid-course. Secondly, the indicators should have the same meaning and interpretation during the ten years, i.e. the progress in the value of the indicator should derive from progress in innovation rather than from changing conditions. An indicator that would improve merely with the passing of time (as, for example, an indicator of adoption of a particular technology) would not do. 6. Internationally comparable. Calibrating the position of Europe relative to other areas of the world is an essential aim of an innovation policy benchmarking exercise. It should therefore be a requirement of the proposed indicators that they be available for other areas of the world, ideally for all areas (emerging countries included) but at the very least for the USA. 7. Decomposable. If not from the launching phase of the Europe 2020 strategy, then at least from very shortly afterwards, the indicators and possible targets should be translatable (in a non-mechanical way) into targets for the different Member States (or even regionally). The adaptability of the indicators to this task will be a key factor in their success since most levers of action reside within the Member States (or regions). A related desirable property is unbiasedness, that is, the indicators should not a-priori be strongly correlated with characteristics that may favour any Member State, or region, over another. This consideration may point towards indicators weighted by GDP (for instance) and targets expressed as rates of change, rather than levels. 8. Low susceptibility to manipulation. An indicator should not measure something so small or so fragile that its very nature and significance could be changed by the behavioural effects induced by using it to set a target. 9. Easy to handle technically. This refers to aspects such as the necessity of consistent measurement at different levels of aggregation. 10. Sensitive to stakeholder's views. In particular, we note that a strong preference for output oriented indicators to complement the R&D intensity 3% target has been expressed. Of course, traits that would make an indicator a-priori unacceptable to a significant segment of stakeholders should be avoided. A caveat is required here. We will often be referring in this report to indicators, benchmarks and targets for the EU as a whole. An overall target has implications for the different stakeholders but these should always take the form of adapted targets. The current situation of each stakeholder under consideration is important. III. INDICATORS IDENTIFIED As already stated, it is the opinion of the HLP that the formulation of indicators that conform to its remit is a complex challenge, for at least three reasons: The first was discussed in Section II. The nature of the indicators considered has to satisfy, if not all, at least a good number of the conditions laid down there, and if not literally at least in spirit; 6

9 Although already included in the presently computable property of Section II let us emphasize, with a sense of some frustration, that data availability seriously constrains what we can recommend; More importantly, we feel that understanding the richness of innovative activity by focusing on a single dimension can hardly do justice to its multi layered character. Consequently we have decided to proceed by, in a first step, presenting a short list of indicators. Each of them, alone, yields only a partial view of the innovation phenomenon, but together they provide if not for a perfect, at least for a clearer assessment. In Section V we discuss two options for a second, sharpening step, although we shall defer on explicitly choosing among them. In Section VI we shall mention some of the data collection efforts that would be necessary in order to go forward. In this section the indicators are presented in a non-technical and general manner. More details can be found in the Technical Annex. FIRST INDICATOR Innovation is an engine of economic growth and, more generally, of the improvement of economic well-being. Therefore, success in the latter should in part be attributed to innovation. Not exclusively, of course. Other factors also come into play. Most notably increases in the traditional factors of production, such as hours worked (we observe that the objectives of The Europe 2020 strategy include an important increase in this factor, through less unemployment and more labour participation) and capital (physical, human, even intangible). It makes sense, however, to conclude that once these have been taken into account, what is left, the residual, must largely be due to the exercise of innovative activity. Unfortunately, the availability of data only allows, in the form of the classical indicator of productivity per hour worked, for a partial measure of this residual. Additionally clouding the matter, productivity per hour worked is quite sensitive to changes due to economic general conditions. It should be recognized therefore that although this indicator undoubtedly provides an output measure - it is in a sense the ultimate output measure - it is nevertheless too indirect and distant from what ideally one would like to reflect, the innovative activity itself, for it to be a candidate for a single indicator that would summarize progress in innovation. Thus, it cannot play for innovation policy an equivalent role to the role played by the R&D intensity indicator for research policy. Yet, the importance attached by the EU 2020 strategy to closing the productivity gap 4 advices to retain it in our first set of indicators, as it is the most closely related to that challenge. Hence, we propose: Hourly labour productivity Given the stated limitations of this first indicator we believe that one needs to concentrate on more proximate, less abstract, measures of the extent of innovative activity, and we proceed to do so. The innovation challenge for Europe is both quantitative and qualitative. We need to innovate more, and we need that our innovative activity be competitive in the world. We shall therefore propose three 4 COM(2010) 2020 final: "What is needed is a strategy to turn the EU into a smart, sustainable and inclusive economy delivering high levels of employment, productivity and social cohesion. This is the Europe 2020 strategy" 7

10 indicators to measure the extent of innovative activity (indicators 2, 3 and 4) and one indicator to control for the quality (indicator 5). SECOND INDICATOR As a starting point we want keep track of the ideas that have crystallized in forms that are ready, in principle, for use as inputs in the productive activities of firms. To accomplish this, the HLP has considered indicators based both on patents and on trademarks. However, trademarks are subject to large cultural and international differences, as well as to variations excessively dependent on economic climate. In consequence we feel that at present they are not suitable as a solid indicator of innovative activity adapted for international comparisons. On the other hand, patents applications are more directly comparable. Moreover, a substantial part of the innovation capacity of a country depends on its ability to transform research and technology development (well approximated by patents) into technological applications. This technological dimension is therefore crucial to assess innovation performance of Europe, although it needs to be completed by other indicators that better cover the non-technological aspects. Hence we propose: Patent applications weighted by GDP 5 THIRD INDICATOR Innovation is applied knowledge. Therefore, the availability of a well educated population is a key asset favouring innovation. The Europe 2020 strategy includes ambitious objectives in this domain and we will not, in consequence, duplicate them here. An educated citizenry is, however not by itself sufficient. At the very least the embedding of education into the economic structure is also a highly relevant aspect. This observation leads us to propose, as the second indicator of the extent of innovative activity, a measure of the size of the knowledge-intensive economic activities. Moreover, the proposed indicator also captures the market demand for innovation, a very pertinent factor favouring the sustainable development of innovation. In defining this indicator the panel took also into account the need to avoid any bias, regarding manufacturing versus services, or technology-oriented versus non-technological innovation. For these reasons we could say that the second indicator of extent that we recommend accounts for the fertile soil on which innovation may develop. Specifically: Percentage of employment in knowledge intensive activities FOURTH INDICATOR As Schumpeter rightly insisted on, intellectual property, a highly skilled population and sizable knowledge-intensive sectors will not by themselves go very far, innovation-wise, without the creative spur of entrepreneurship. We need to supplement the two previous measures by an indicator that allows us to calibrate the presence, the successful development and the dynamism of innovative entrepreneurial activities. This indicator would measure, so to speak, the progress in 5 The ratio of patent filings to GDP relates patent activities to economic activity and improves comparability across countries, independently of country size and R&D specialisation. This ratio is used by WIPO 8

11 the conditions for quick change, thus ensuring that Europe is on the move, rapidly enough and with the right engine. When we analyse data, we observe a striking, world-wide phenomenon: successful economies are characterized by the emergence of fast growing firms. This is so in traditional industrialised countries such as the USA, but also in emerging countries. With the aim of formulating a specific indicator for this phenomenon the HLP examined several sources of data. The aim was to identify the enterprises that, first, are fast growing and, second, do so in knowledge intensive activities, as those are the key instruments for the acceleration of the structural changes and the adaptation to societal needs that Europe requires. Unfortunately, it appears that at present we have to face the unpleasant fact that not enough data exists from which an adequate indicator could be constructed. As a temporary solution, the possibility of using investment of SMEs in R&D was discussed. Ultimately, however, this indicator was not retained: it was not felt sufficiently representative of the phenomenon that we want to capture, and it has an excessive overlap with the headline R&D indicator of the Europe 2020 strategy. It follows from the above that we need to consider as a matter of extreme priority the development of more comprehensive and inclusive data collections in this domain. Based on the work of Eurostat and OECD in the "Entrepreneurship Indicators Programme", as well as on the work of the Directorate General RESEARCH of the European Commission and Eurostat on Knowledge Intensive Activities, the judgement of the HLP is that the development of such an indicator can be reasonably expected within two years, provided sufficient political back-up and resources are made available. Under such an hypothesis the indicator may have the potential to constitute an Innovation Headline Indicator. We elaborate on this in Section V. For the record we state in general terms (and some uncertainty) the fourth recommended indicator is therefore: Share of fast growing (or young?) and innovative firms in the economy FIFTH INDICATOR We contend that the combination of indicators 2, 3 and 4 offers us a reasonable summary of the extent of innovative capabilities. It tells us something significant about the amount, actual or potential, of innovation. It does not, however, tackle the issue of the quality of innovation, which is also paramount. Therefore, to complete the picture we need an indicator that addresses this aspect. We focus on indicators of the competitiveness of the knowledge-intensive sectors, as measured by the balance of trade. We tested several methodological approaches. Faced with severe data limitations we resort to an approach tried by the OECD as regards the exports of high tech and medium-high tech manufactured goods (in the benchmarking exercise of the next section we also discuss services). In summary: Contribution of innovative-related trade in manufactured goods to the balance of trade of goods 9

12 IV. BENCHMARKING EUROPE AND THE MEMBER STATES In this section we first benchmark the European Union in relation to the US, Japan and South Korea, then we present an analysis of each indicator at the level of the member states, and finally we gather some remarks on targets. 1. CLOSING THE GAP: BENCHMARKING THE EU WITH THE US, JAPAN AND SOUTH KOREA The following graphs present the performance of Europe in the aggregate as compared to other developed economies for the four indicators for which a satisfactory definition, and corresponding data, is available. These are Indicators 1, 2, 3 and 5 (for Indicator 3 data is missing for South Korea). Indicator 1 Labour productivity - GDP per hour worked in US$, 2009 Indicator 2 PCT Patent Intensity (2), 2007 EU-27 (1) US EU-27 US Japan Japan US$ South Korea South Korea Indicator 3 Employment in Knowledge Intensive Activities (KIA) (excluding the public sector) as % of total employment, 2009 (4) Figure 4 Contribution of high-tech and medium-high-tech manufactured goods to the trade balance, 2008 EU-27 US EU-27 (1) US Japan Japan South Korea % % Source: DG Research Data: Eurostat, DG ECFIN, OECD Notes: (1) EU-27 does not include BG, EE, CY, LV, LT, MT, RO, SI. (2) Patent applications under the PCT (Patent Cooperation Treaty), at international phase, designating the EPO by country of residence of the inventor(s) per bilion GDP(PPS ). (3) EU-27 does not include BE, IE, EL, IT, LU. (4) JP: 2005; US: As for Indicator 4 (Share of fast growing and innovative firms in the economy) we find ourselves in a difficult position. On the one hand this is arguably a fundamental 10

13 indicator, even with the potential to become a headline indicator. But on the other, and given current data availability, there is not at present the possibility to come up with a satisfactory precise indicator. As mentioned in the previous section, we estimate that with a determined effort a solid indicator could be available in two years time. 2. ANALYSIS BY INDICATOR 2.1 Indicator 1 Hourly Labour Productivity Figure 1 Labour Productivity - annual growth rate 6 5 South Korea 4 % US Japan EU-27 (1) Source: DG Research Data: OECD Note: (1) EU-27 does not include BG, EE, CY, LV, LT, MT, RO, SI. A natural aspiration for Europe would to aim at a level of productivity growth similar to the USA. 11

14 Figure 2 Labour Productivity - GDP per hour worked in US$, 2009 Luxembourg Ireland US Belgium Netherlands France Germany Sweden Austria UK Spain Denmark Finland EU-27 (1) Italy Japan Greece Slovenia Slovakia Portugal Czech Republic Hungary South Korea Estonia Poland US$ Source: DG Research Data: OECD Note: (1) EU-27 does not include BG, EE, CY, LV, LT, MT, RO, S Note that an increase of 10 points (approximately 20%) would bring the European average value of the indicator roughly to the level corresponding to France in Of course, this is to be added to the aspiration associated with the trend growth reflected in Figure 1. 12

15 2.2 Indicator 2: Patent applications weighted by GDP Figure 3 Evolution of PCT patent Intensity (1) 8 7 Japan 6 South Korea US EU Source: DG Research Data: Eurostat, DG ECFIN, OECD Note: (1) Patent applications under the PCT (Patent Cooperation Treaty) per billion GDP (PPS ). The significant gap between the European Union and Japan and South Korea is striking. Note that the progress by these two countries took place in a rather short period of time. In the view of the panel this shows that there is room for rapid improvement in Europe as well. 13

16 Figure 4 PCT Patent Intensity (1), 2007 Sweden Finland Denm ark Germ any South Korea Japan Netherlands Austria US EU-27 France Belgium UK Ireland Slovenia Italy Es tonia Hungary Spain Luxembourg Malta Czech Republic Latvia Portugal Slovakia Bulgaria Greece Lithuania Cyprus Poland Rom ania Source: DG Research Data: Eurostat, DG ECFIN, OECD Note: (1) Patent applications under the PCT (Patent Cooperation Treaty) per billion GDP (PPS ). 14

17 A cross-country correlation analysis shows that there is a connection between Patent Intensity and R&D investment. However, the correlation is far from strict, thus demonstrating that this indicator is related to innovation capacity in terms of the ability of a country to transform efficiently research into protected knowledge. For instance, The Netherlands which has a far lower R&D intensity than Japan, perform nearly as well as Japan in terms of Patent Intensity. Other countries such as Ireland and Slovenia show the same pattern of good performance in terms of Patent Intensity in spite of having more modest level of R&D intensity. Note that, other things equal, an increase of 2 points (approximately 50%) would bring the European average value of the indicator roughly to the level corresponding to The Netherlands in

18 2.3 Indicator 3 Percentage of employment in knowledge intensive activities Figure 5 Share of employment in Knowledge Intensive Activities (KIA) (excluding the public sector) in total employment, 2008 and 2009 Luxembourg Ireland UK Sw eden Finland Denm ark Netherlands Germ any Belgium Malta France Austria US EU-27 Italy Cyprus Japan Slovenia Hungary Spain Czech Republic Gree ce Es tonia Slovakia Lithuania Latvia Poland Portugal Bulgaria Romania % 2008 (1) 2009 (2) Source: DG Research Data: EUROSTAT, OECD 16

19 This indicator constitutes a macro-economic variable and therefore has a strong economic weight. Surprisingly, it shows rather significant variations from one year to another (for example, for Slovenia, Ireland or Estonia). Therefore, it makes sense to expect that a far-reaching economic strategy, such as the Europe 2020 strategy, could influence the design of policies making possible its significant evolution, hence of the capacity of Europe to innovate. It is interesting to note that countries which traditionally perform very well in the EU innovation scoreboard index, such as Finland, Sweden, Germany or Denmark, continue to perform extremely well in terms of Indicator 3, but are challenged, or even outperformed, by other countries such as Luxembourg, Ireland, the UK or the Netherlands, Malta and Austria. Note that, other things equal, an increase of 2 points (approximately 10%) would bring the European average value of the indicator to the level corresponding to Germany in

20 Indicator 5 - Contribution of trade in innovative-related goods and services to the balance in trade of goods and services Goods Figure 6 Contribution of high-tech and medium-high-tech manufactured goods to the trade balance, 2000 and 2008 De nmark Japan UK Germany Hungary France Sweden Ireland Italy Belgium EU-27 (1) US South Korea Slovakia Austria Finland Spain Portugal Greece Czech Republic Netherlands Poland Luxembourg % Source: DG Research Data: OECD (1) EU-27 does not include BG, EE, CY, LV, LT, MT, RO, SI. 18

21 This indicator uses an approach tested by the OECD as regards the trade in goods. It calculates the contribution of some categories of goods to the overall trade balance. It constitutes a good indicator of specialisation in international trade for innovation-related goods. Moreover, the link to the trade balance addresses the issue of imports and re-exports after improvements on the goods. The results demonstrate the potential of this indicator to embrace a rather large concept of innovation, as shown by the good performances of the Slovak Republic and of Hungary. Denmark is the best performer, with Japan in second place. Nearly all the European Member States registered significant progress between 2000 and 2008, old Member States such as Denmark, Austria or Finland, as well as new ones such as Slovakia and Poland. Note that, other things equal, an increase of 2 points (approximately 40%) would bring the European average value of the indicator roughly to the level corresponding to Germany in

22 2.5.2 Services Figure 7 Contribution of KIS to the trade balance, 2002 and 2007 Norway Denmark Greece Ireland Germany Luxembourg Netherlands Latvia Japan Cyprus Romania Hungary Czech Republic Estonia Portugal Sweden Poland Italy Slovakia Lithuania US Spain Slovenia Bulgaria France Finland Austria Malta % Source: DG Research Data: UN COMTRADE, OECD, EUROSTAT 20

23 The Panel applied the same methodology to the trade in knowledge-intensive services (KIS). But here the result has to be regarded as less conclusive, at least for the moment (data improvement efforts are on-going and at an advanced stage). On the one hand, while in the case of manufacturing goods the groups of products studied are large, in the case of knowledge intensive sectors, the composition of this group is more limited, relying heavily on financial services or transport of oil and gas. Performances may therefore depend to a large extent on the specific factors affecting these sectors (for instance in the case of transport the geographical situation of the country, e.g. Norway, or in the case of financial services the tax treatment of transactions). In the case of the USA, the negative performance may, at least partially, be accounted for by strategies of fiscal optimisation. 3. TARGETS The issue of targets can be discussed in reference to the EU-27 globally considered or in reference to the different Member States. Targets for the EU-27 globally considered The present report does not propose targets. Here we have carried out the preliminary work of identification of suitable indicators and of benchmarking them in relation to other developed economies in the world. Of course, a next logical steps could be setting targets for the EU-27 globally considered (in terms, for example, of percentage improvements in the different indicators). A judgement on the appropriateness of doing so, as well as the specific details, is a matter under the responsibility of the political decision-makers, but we hope that the benchmarking exercise carried out may prove useful and perhaps suggestive. It is important to realize that for the setting of ambitious but realistic targets the political decisionmakers will require a good understanding of the mechanism by which policy can influence outcomes. This is indeed indispensable in order to have a reasonable assessment of the policy changes (in areas such as risk-capital, patent law, public procurement regulations, tax incentives, etc) that should be promoted from, or at, the EU level if the targets are to be attained. Targets for the Member States Targets for the Member States could derive, in part, from the direct adaptation of global European targets (if and when in place) and, in part, from internal benchmarking in Europe (that is, in relation to other Member States). Because of this second factor the nature of the exercise may be simpler for the Member States. Still, two relevant considerations are: The targets (even if they are formulated not on the level but on the rate of change of an indicator) cannot, and should not, be the same for every Member State. This is because a unique model for an innovation development path does not exist. The specific structure of the country s economy, not only in terms of economic sectors, but also in terms of type of firms and inter-relations with the public sector, matters. It would therefore be difficult to propose a uniform strategy, reflected on uniform targets, for increasing Europe's innovation capability and innovation efficiency. The targets of the Member States should be inspired by the global targets (if they are present) and by the example of other Member States, but also adapted to its own characteristics. If global targets are formulated, the targets of the different Member States should fit together and be demanding enough to yield a good chance of 21

24 attainment of the global targets. Coordination in the adaptation of the latter by the different Member States is therefore indispensable. If, as contemplated in Option B of the next Section, a single headline indicator was to be adopted, the fixing of a target at the EU-27 level would be a simpler matter (assuming, of course, that the indicator is available on a statistically sound basis). Yet, there would be a loss of flexibility at the Member States-level since the degrees of freedom for the local adaption would be reduced. In view of this, the HLP points to: The importance of testing the actual data before speculating on an adequate level for the Member States indicator, The necessity to consider targets in relative, rather than in absolute, terms (percentage of growth, etc.), as there may be too much uncertainty about the optimal absolute levels associated to specific national, economic and technological conditions. V. OPERATIONAL SCENARIOS The HLP had in depth discussions on how to strike a balance between the desirability of coming up with an immediately operational indicator and the need to be sure of the merits and quality of such an indicator. The desirable properties for the latter spelt out in Section II bear witness to the difficulties of such an exercise. The judgement of the HLP is that at this stage we cannot offer solid arguments to narrow down the list of suggested indicators to a single one among those that would be immediately available. In this Section we offer two options (Option A and Option B) to approach the fulfilment of the remit of the HLG. We believe that the choice among the two options goes beyond the strictly technical and that it should be made at a more political level. Procedurally, we shall present each of them in the best possible light. 1. OPTION A: A LIST OF THREE INDICATORS Our list of five indicators could be reduced to a triad by retaining those immediately available and more directly related to the role of knowledge. That is, by dropping the indicator on fast growing innovative firms and on productivity. A cursory correlation analysis makes us believe also that retaining the productivity indicator would not add substantially to the combination of the other three. We would in this manner have two indicators related, in different ways, to the measurement of the extension of the soil most conducive to innovation and a third indicator related to the competitiveness of the latter. We also point out that as it has been done in the case of climate change (with the 20/20/20 target) it may be possible to formulate a suitable and appealingly simple triadic target (in the previous section we included some suggestions). It is appropriate to mention that the HLP discussed the possibility to develop a composite indicator based on the short list of indicators proposed in Section III, or even in the three that we are now retaining. Because of the reasons expressed in Section II, as well as the methodological weaknesses caused by the structural differences among Member States, this was considered unadvisable. The HLP therefore confirms a preference for a short list over a composite. In case the political decision makers that hold the final responsibility were to opt for a composite, the HLP would like to express the view that the number of components 22

25 in such a composite indicator should be limited, so as to ensure: (i) the necessary transparency of the indicator, (ii) the relative independence of those components, (c) the possibility to monitor adequately the policies addressing the various components. 2. OPTION B: FAST GROWING INNOVATIVE FIRMS As already mentioned in previous sections the HLP identified as an interesting indicator, and as a promising candidate to a headline indicator, the share of fast growing innovative firms in the economy (Indicator 3 of Section III). This is indeed a measure of entrepreneurial activity which is directly linked to innovation and that can be readily understood and communicated to a large public. Option B proposes to focus on this indicator. Doing so would imply a conviction that this is a key factor for the vitality of innovation and that Europe has weaknesses in it. Both are defensible propositions. The evidence from the USA on the role of fast growing innovative firms on economic performance is substantial 6. They have been a driving force in the development of emerging industries and, in a virtuous circle, they have been boosted by the numerous opportunities generated by these industries. Moreover, spin-offs from universities and research centers have been a highly effective means of insuring the transfer of fundamental knowledge from its birthplace to a downstream development site. Although data are sparse and partial, it seems that, in comparison, Europe is characterized by a strong deficit in term of fast growing innovative firms. A dynamic business sector is at the hearth of growth, creativity and innovation. In this respect the indicator suggested would be forward looking and compelling (young innovative firms need to grow to create employment), mobilizing, (it stresses the role of business in innovation), analytically very relevant, and with strong links to policy. In addition, it would be an integrative indicator: without being a composite, it would summarize many of the relevant dimensions of an innovation system. To wit: Framework economic conditions (financial markets and access to credit, education, orientation of economic institutions towards entrepreneurship and dynamism). Structural changes. The emergence of a population of fast growing innovative firms is a sure sign that something disruptive, and positive, is happening. Likelihood that a good representation of the next generation of very large firms, leaders in their domains, originate in Europe. Last but not least, the indicator would also constitute a good complement of the R&D intensity indicator. The latter tries to capture the general progress of an economy towards innovation (and it has a significant public component) while the former focuses on a single and powerful driver for innovation and growth (and stresses the role of business). 6 See, for example, Haltiwanger, Jardin and Miranda, 2009,: Business Dynamics Statistics Briefing: Jobs Created from Business Star-ups in the United States, Erwin Marion Kauffman Foundation 23

26 Settling on Option B would require the immediate launch of a two-year program of work aimed at the precise formulation of the indicator as well as to the statistical, data gathering, and even legal, efforts associated with it. Europe needs, as matter of priority, better information concerning ambitious and innovative firms able to grow in innovative markets. Current data collection efforts do not provide an amount of data on growth rates and domain of activity that would allow the clear identification of a statistical group of Innovative fast growing firms. This information should be collected for firms with patent activities (SMEs which patent, by fields of activities) as well as for firms which do not patent but are growing in knowledge intensive activities. We therefore strongly recommend a European Commission program of work for the next two years aiming at designing and compiling the relevant indicator(s) and data sets. We expect that cooperation between Eurostat and OECD, in the form of an enhanced version of the ongoing EIP (Entrepreneurship Indicators Program), would allow delivery of the needed indicators in the specified time frame. VI. FORWARD The extremely limited time available for the production of this report 7 did not allow for an in-depth analysis of all the issues involved in the measurement of innovation. However, the dialogue within the Panel, as well as the substantial contributions received from the Member States, from the OECD and from the Commission services, made clear that Europe needs to improve the understanding of the role of innovation in its economic performance, and thus the HLP calls for a substantial research effort on conceptual foundations, mechanism of policy transmission for the case of innovation, as well as on empirical testing and measurement. In the same light, the HLP views as a crucial challenge for Europe, but also for third countries, the need to work together to improve the collection, treatment and analysis of some of the key data related to innovation. Some of the more specific issues have been mentioned already in previous sections. The following is a non-exhaustive summary of some of the key areas where improvement is needed in the short and the medium term. 1. PRIORITY ISSUES Two areas have been identified by the Members of the panel as key candidates for urgent development. Fast growing innovative firms Measuring the share in the economy of such firms is an issue of utmost priority. In all likelihood doing this properly will require legislative action at EU level. The availability of a high-quality and comprehensive data infrastructure (including the regional level) is crucial. The backbone of such infrastructure is a high quality business register. Currently we have EU Regulation (EC) No 177/2008 which, for statistical purposes, establishes a common framework for business registers. But to secure extensive coverage, methodological harmonisation and timely provision of good quality data, further legislative development is needed. 7 The period of activity of the HLP was the summer of The Panel met three times. There were also many interactions at a distance. The help of the Commission services was intense and invaluable 24

27 Innovation and productivity Innovation should ultimately be reflected on labour productivity. It is, however, only one of the numerous factors affecting it (others are physical and human capital, intangible forms of capital, even social capital). A better understanding of the interactions between innovation and labour productivity is necessary to apprehend and effectively measure the share of the changes in productivity related to the investments in innovation. In particular, focused research, and data gathering, should urgently be carried out on total factor productivity, a concept which can best account for the share of innovation in the changes in productivity at the global level. 2. OTHER ISSUES FOR FUTURE DEVELOPMENT Some of the other key areas where improvement is needed in the short and medium term are: Enhancing the coverage, quality and informative value of proposed indicators Knowledge intensive activities International data based on the EU definition of Knowledge Intensive Activities (KIAs) is urgently required in order to provide European policy makers with relevant comparisons. An international definition of KIAs would also help. More precision in data (in more technical terms: data at a more disaggregated level than level 2) would also be welcome in order to better understand sub-sectors in term of their "knowledge intensity". Knowledge intensive services and innovation Services, in particular knowledge intensive services, are a key and growing component of the economy. The identification of the precise domains that are linked to innovation (as users and as providers) is needed in order to put forward a specific classification for innovation-related services. The development of measurement methodologies free of obvious biases is also a challenge, as is harmonization, i.e. ensuring coherence and compatibility of data between various systems (OECD, UN, Eurostat, ) As an example, the notion of medium-high tech goods is not shared between OECD and Eurostat. Advancing towards a consensus in the formulation of a composite indicator of balance of trade in innovative goods and services is also important. Improving current systems for measuring innovation Community Innovation Survey (CIS) Some of the needs are: Refinement and redevelopment of the CIS to ensure that the information is more reliable and robust, including using more standardized methodologies across countries within Europe and with third countries. Validity tests of innovations stated by respondents and links between CIS data and other data sources. There is also a task of building up a consensus on an adequate balance between the quality of the data and their periodicity, for example, on the categories of data for which the periodicity should be annual, and those needed only every two or three years, but with enhanced quality. European Research and Innovation scoreboard The future European Research Innovation scoreboard could provide a foundation for the launching of an EU innovation Index providing an "EU wide Innovation capability 25

28 Measure" grounded on a composite of performance indicators for innovation. It is important, however, that it be based only on the components of the scoreboard that are internationally comparable and that the weight of the different components enjoys a sound economic rationale. Medium term issues a): Measurement of new forms of innovation Open, distributed and user-driven innovation Innovation increasingly involves interactions among different actors in diverse ecosystems that span a range of different organizations. As of now, we have only weak or partial measures of these types of interactions. Tools should be developed for the measurement of firm-level and population-level data on open innovation, networks, alliances and user-involvement in innovation. Innovation in the public sector To date our focus on innovation has been mainly concentrated on private sector innovation or on the outputs of the research system. Although valuable, this information says little about innovation in the public sector. Greater efforts are required to extend and enhance our understanding and measurement of innovation in the public sector. Creative industries Europe has a strong and dynamic creative industry sector and the contribution of this sector to economic development should be more fully assessed and measured. Inclusive, sustainable and smart growth including Green or Eco-innovation We still lack indicator measures and data on European progress in "inclusive, sustainable and smart growth", green or eco-innovation. This is a critical area of future development to help achieve the 2020 ambition. Intangibles and Innovation Current measures of innovation cover only part of firms' investment in innovation. A wider set of measures is required that capture the full range of investments in intangibles, and that allow to test their possible contribution to innovation, productivity and growth. In particular, research on the impact of trademarks and on how to make possible its international comparability would be most useful. Medium term issues b): Further work on the understanding of innovation Taxonomy of Innovation Reaching an agreement on an overarching definition of innovation, and an associated taxonomy, is important in determining the utility and universality of innovation indicators. Making better use of Innovation indicators and statistics Linking EU Innovation strategy effectively to innovation indicators In this Report the HLP has explored a (limited) number of indicators. The issue of targets based on them has been addressed only in part, as it is primarily a policy issue. However, the rationale for fixing targets at EU level, the economic analysis of its impact and the criteria that could be used by Member States and regions to set corresponding national or regional targets, are all issues on which economic analysis and views from the business sectors and other stakeholders would be helpful in 26

29 order to identify how best to connect the indicators to the emergent EU Innovation strategy, and also to asses how the indicator(s) might be used to track progress. Better diffusion tools More interactive / user friendly tools could be designed in order for the EU innovation policies to reach the wider public. 27

30 TECHNICAL ANNEX TO SECTION III 1 - Labour productivity Definition: GDP per hour worked. Labour input is defined as total hours worked of all persons employed. Source: OECD Geographical coverage: Covers 19 EU Member States: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, UK, Czech Republic, Hungary, Poland, Slovakia. The OECD data allows for a comparison between major world economies but does not cover the whole of the European Union. A detailed table on labour productivity containing data for most Member States is available from EUROSTAT but does not provide the same elements for international comparison available from OECD.. Availability: Available with a two years delay Frequency: updated yearly 2- Patent applications weighted by GDP Definition: PCT patent Applications per billion GDP (PPS ) Source: WIPO, Eurostat, DG RTD, DG ECFIN, OECD Availability: Available with a two years delay Geographical coverage: large and consistent international coverage (WIPO) Frequency: updated yearly 3- Percentage of employment in knowledge-intensive activities Definition: Knowledge-intensive activities are defined as economic activities where persons with an academic oriented tertiary education (ISCED 5 and 6) represent more than a given percentage 8 of the employment, to the exclusion of the following activities where the service provider is predominantly the public sector: 8 A tentative percentage of 33% has been used for the purpose of this report. 28