Understanding Technology Specialization and Its Importance in the EU Economy, Lecture notes of Technology

Download Understanding Technology Specialization and Its Importance in the EU Economy and more Lecture notes Technology in PDF only on Docsity! 1 Technology and Specialization: Dilemmas, Options and Risks? Tassos Giannitsis1, Marianne Kager2 Expert group “Knowledge for Growth”, May 20093 1 University of Athens, Department of Economics. 2 Economist. 3 The views expressed in this paper are purely those of the authors and may not in any circumstances be regarded as stating an official position of the European Commission. 2 CONTENTS 1. Introductory remarks 2. Specialization and specialization policies: basic concepts and data 3. Why does specialization matters? 4. Measures of specialization 5. R&D Structure and specialization patterns: the landscape in the EU 5.1. A general view of R&D specialization in the technological advanced E.U. countries 5.2. Technology intensity of trade specialization of 8 E.U. member countries 5.3. The EU R&D regional landscape 5.3.1. R&D intensities on NUTS 1 level and specification patterns 5.3.2. Regional typologies of RIT system 5.3.3. Summary 6. Technological-led specialization as a wider E.U. policy issue 6.1. The issue of appropriate policy mixes 6.2. The risks from neutral policies 6.3. The risks of targeting specialization 6.4. An horizontal issue: Concentration of resources versus decentralization approaches 6.5. Three different strategies 6.5.1. Strategies aiming at the frontier 6.5.2. Preventive strategies to face global risks 6.5.3. The catching-up and the convergence issue 7. A lesson from the financial crisis: The importance of regulating risky innovation- based specializations 8. What choices and how to deal with risks? 9. Bibliography 5 of knowledge. What matters is the ways new knowledge and technical change are integrated and assimilated in productive processes and with what results. Such a view allows a widening of the context in which technology policies are analyzed, that is the creation of new scientific knowledge, new technologies, innovations and the development of new dominant and emerging technology areas. Moreover, it is important to consider that new knowledge and technology are not related to specialization in a linear way, that the possible combinations of knowledge and technologies and their future uses and implications are unpredicted and can lead to competitive advantages of very different nature and in very different environments. This means that the point at issue is not only technological. It is of a more complex nature, linking effective governance, coordination of research and technology policy, knowledge building and the shaping of productive processes. Specializations are not constructed only from knowledge inputs, related to technological change and innovations, leading to new products and/or processes. Concurrently to these factors, it is important to consider the «classical functions» of production processes, in which technological inputs are associated with very diverse locally available labour, capital, other inputs and, in particular, the prices of these inputs in different combinations. In such a framework, the ‘game’ of who can create competitive positions becomes complicated. The same knowledge permits producing firms to achieve very different combinations of the various elements of the value chain and hence to construct “differences” and competitive advantages through several specialization forms. Early-entrants are often advantaged. However, in the process of time, imitators can also be successful, even vis-a-vis the original innovators, to the extent that through a smart use on the one hand of these new technologies and on the other hand of their production value chain, they succeed to create specific or niche competitive advantages. In such cases, different rational behaviors lead to different efficiencies and might open windows of opportunities resulting to a change of previous hierarchies of competitive advantages. Such a differentiating attitude allows also a more flexible approach to policy issues regarding leading players and followers. Obviously, technologically leading countries and firms are the environment in which most evolutionary or more radical technical change develop. These actors possess the research capabilities to generate new technology paths and to better benefit from such policies. In contrast, followers and laggards may be disadvantaged in all dominant and emerging technologies. What could be the policy consequence? That the efforts of laggard countries and firms to catch-up in these areas are condemned to lead to inefficiencies and a waste of resources? That there is no scope for policies enhancing advanced technologies in view of catching-up? A positive answer would be very deterministic, and history is plenty of examples of followers who were quite successful in imitation or adaptation strategies. Besides, all the significant efforts of the EU and national governments to diminish the gap to technology leaders in ICT or other areas indicate that specialization policies matter. Comparisons of specialization patterns and competitive advantages between leaders and laggards need to be regarded from different points of view. For example, even within these two groups very diverse research and technology capabilities and specialization characteristics can be observed. In fact, the EU members, according to their R&D intensity (GERD as % of GDP, 2006) can be classified in four main clusters (or “clubs”): - above 2,5% (Sweden, Finland, Austria, Germany), 6 - 1,5% to 2,5% (Denmark, France, Belgium, UK, Netherlands, Slovenia, Czech Republic), - 1% to 1,5% (Hungary, Italy, Estonia, Spain, Ireland, Luxembourg), and - below 1% (Portugal, Lithuania, Latvia, Greece, Poland, Malta, Bulgaria, Slovakia, Romania, Cyprus)5. To restrict the comparison between followers and leaders in the aim to draw policy conclusions on specialization approaches would be insufficient. Technological capabilities, specializations and competitive advantages should not be categorized in dichotomous terms such as leaders-followers or strong-weak. As useful this distinction can be for other reasons, in this case it would disregard the fact that the concepts of strong and weak are relative ones. Specialization and competitive advantages of a country or firm can at the same time be weak and strong. Weak vis-à-vis the upper side of the pyramid and strong vis-à-vis other weaker followers. A different view results when we compare specialization patterns and competitive positions of followers to the “upper cluster” than to countries and firms with inadequate specializations and weaker competitive positions. Finally, it can be argued that specialization is not only the result of the creation of knowledge and/or technologies, but also of the diffusion of such technologies and the capabilities of firms and countries to exploit them and to use them in different applications or in a broader productive scope. In all these cases we are faced with forms of imitation or adaptation or with complementary strategies, which create specializations even in apparently similar fields, but with diversified characteristics, leading to the pursued competitive “differences”. A very large part of countries and firms in the international scene have been followers or laggards, but nevertheless they could advance through technology transfer, adaptation and imitation. 4. Measures of specialization In the literature, the measurement of specialization originates in trade theory. A variety of specialization indices have been developed to capture the specialization of a country and various such indicators have been also used as technology specialization indices, of course after some adjustments. The most common technology specialization indices have the following form: Concentration indices, measuring the weight of the n more important sectors (n can take the value of 1, 3, 4, 6, etc) to the total relevant figure for a specific technology variable (e.g. R&D, patents). The concentration ratio is given by the following expression: 1 1 ( ) ( ) x j k j n ij i p k C X p = = = ∑ ∑ 5 See also the distinctions in ch. 5.3.2. 7 where X is the number of the largest classes (e.g. level of 4 or 6 classes) to take into account, ijp is the number of patents for application of country j in the ith technological class and ijp is the kth largest number of patents per technological class6. b) Indices very closely related to the Revealed Comparative Advantage Index of the Trade Theory, taking the form: - The Balassa Revealed Technological Advantage Index7 (or specialization index) (SI) defined as following: 100* tanhln / /kj kj kj kj kj j k kj SI X X X X ⎡ ⎤⎛ ⎞ ⎛ ⎞ = ⎢ ⎥⎜ ⎟ ⎜ ⎟ ⎢ ⎥⎝ ⎠ ⎝ ⎠⎣ ⎦ ∑ ∑ ∑ where kjX can indicate alternatively the number of patent applications, exports, etc. of country k in sector j. Positive values mean that the sector has a higher weight in the portfolio of the country than its weight in the area of comparison (e.g. EU). Negative values indicate specialization below the average. - Alternatively: an index of technological specialization8: ij j HH i j j H X MS X ∈ ∈ = ∑ ∑ ij j LL i j j L X MS X ∈ ∈ = ∑ ∑ H i i L i MSITS MS = The index shows how much any particular country or region adapts its relative high to low- tech products trade structure to changing patterns of world trade in high and low technology products. In the above formula i stands for countries (or regions), j stands for SITC products, MS stands for share in the world market, H is the set of high-technology product groups, L is the set of low-technology product groups, ijX is the value of exports to the world from country i in product group j, jX is the value of exports to the world from all countries in product groups j. Comparing ITS in different time periods shows how a country’s relative market share in high and low technology change. A value below (above) one indicates that a country’s 6 See N. van Zeebrock et al., (2006), Issues in measuring the degree of technological specialization with patent data, Scientometrics, 66, Nr.3, p. 484. 7 See Erawatch, (2006), R&D Specialisation. Methodology and Data used. 8 L. Alcorta, W. Peres, (1998), Innovation systems and technological specialization in Latin America and the Caribbean, Research Policy, 26, pp.873. 10 Second, while the three Mediterranean countries are all characterised by similar weak competitiveness with regard to the medium to high-tech industries, there are large divergencies regarding the competitive performance of the most advanced countries. Third, as a result of teir weak position in medium and high tech industries, the technological less advanced countries suffer from overall trade deficits, since their technological disadvantages cannot be counterbalanced by stronger competitive positions in the technological less intensive export categories. Table 2: Technology intensity of the export structure (2006) Products Greece Spain Portugal Ireland Germany France UK Netherlands Agricultural products & Raw material 10,8% 6,6% 5,9% 9,0% 4,4% 7,6% 3,9% 12,3% Low technology 24,3% 19,9% 24,4% 8,8% 11,9% 15,3% 11,6% 13,6% Low- Medium technology 40,3% 23,5% 30,1% 2,6% 17,5% 17,7% 21,9% 23,9% Medium- High Technology 14,1% 42,3% 31,3% 42,4% 53,3% 47,9% 35,7% 27,8% High Technology 7,9% 6,5% 6,0% 33,6% 9,7% 9,8% 22,9% 17,7% TOTAL 100,0% 100,0% 100,0% 100,0% 100,0% 100,0% 100,0% 100,0% Source: T. Giannitsis et al., Competitiveness and Technology in Greece (2008, in Greek), Calculations based on Eurostat data. Table 3: Technology intensity of the import structure (2006) 11 Products Greece Spain Portugal Ireland Germany France UK Netherlands Agricultural products & Raw material 8,9% 9,2% 10,3% 6,6% 7,1% 6,3% 6,4% 9,1% Low technology 14,6% 12,5% 16,2% 14,3% 13,2% 14,6% 14,9% 13,2% Low- Medium technology 34,6% 31,1% 31,4% 18,6% 28,3% 30,6% 22,6% 30,5% Medium- High Technology 30,4% 37,1% 32,1% 30,5% 37,1% 37,1% 29,7% 26,4% Source: As in Table 2. Table 4 : The Balassa Index of Specialization by technology groups for EU countries, 2006 Products Greece Spain Portu- gal Ireland Germa- ny France UK Nether -lands Sweden Agricultural products & Raw material -0,428 -0,366 -0,455 0,360 -0,142 0,046 -0,377 0,203 -0,159 Low technology -0,291 0,015 -0,010 -0,025 0,048 -0,024 -0,264 0,070 0,198 Low- Medium technology -0,443 -0,342 -0,232 -0,639 - 0,140 -0,311 -0,159 -0,065 -0,086 Medium- High Technology -0,734 -0,149 -0,223 0,366 0,274 0,080 -0,055 0,080 0,112 High Technology -0,619 -0,388 -0,391 0,377 -0,026 -0,100 0,073 -0,005 0,195 TOTAL -0,502 -0,213 -0,211 0,215 0,099 -0,048 -0,145 0,055 0,076 Source: As in table 2. 5.3. The E.U. R&D regional landscape 12 In the recent past some several studies about regional R&D systems11, R&D intensities as well as about regional R&D specification have been published. With respect to possible future policy considerations it seems beneficial to recall some of the outputs of these studies. 5.3.1. R&D intensities on NUTS 1 level and specification patterns With regard to this level of regional classification three categories of specialization patterns have been distinguished: a) Regional R&D intensities; b) Scientific specializations, and c) Technological specializations. 5.3.1.1. Regional R&D Intensities Overall GERD: • As expected R&D Intensities show high variances between the regions. They differ between 4,16 GERD in % of GDP for Baden-Würtenberg/ GE and 0,12% in Severna/BG. at NUTS 1 level and an 8,7% (Braunschweig /GE) and 0,1% (Swietokrzyski/PL) at NUTS 2 level12, • Out of 79 regions only 5 regions have achieved the Lisbon goal of at least 3% of GERD as of 2005. And even worse, only 17 have an R&D intensity of more than 2%, while 29 regions show a R&D intensity of less than 1%. BERD • The figures show that the regions with the highest BERD expenditure are concentrated only in a few Member States. • Even if one extends the list of regions to all regions which perform a BERD just above EU average, they are concentrated in only nine Member States , ten in Germany, six in France, four in UK two in Austria, Belgium and Spain, and each one for Italy, Finland and the Netherlands. • In most of the TOP performing regions the R&D expenditures are financed and controlled by the business sector. The main exception is Berlin, which has an extraordinary high public expenditure (probably because of the Max Planck Gesellschaft), a fact which is certainly connected with the post war history of Berlin. 11 Fraunhofer Gesellschaft (ISI) recently published two studies: H. Kroll, Th. Stahlecker “Regional Key Figures 1-2008; V. Peter, R.Frietsch:”Exploring regional technology specification; 12 Tables and the Annex are omitted. 15 bordering to London, the state of Denmark and prosperous urban regions in Germany and France. Type 1b: Business oriented R&D driven regions. Very high patenting and publishing intensity: Among them German regions like Tübingen, or cities like Karlsruhe, Scandinavian regions like Stockholm, and Southern Finland. Type 2: Public sector centered R&D supported regions. Very high publishing and slightly over average patenting intensity: mainly capital regions (high public expenditure) like Prague, Bratislava, London and some university towns in Austria, Germany Netherlands and the UK. Type3a: R&D supported regions, comparatively public sector oriented. Average publishing patenting and publishing intensity, but slightly higher publishing and somehow lower patenting intensity. Type 3b: R&D supported regions, comparatively business oriented. With lower publishing and higher patenting intensity. Type 4a: R&D lagging behind, public sector oriented. Near absence of patenting, very low business sector intensity: new member states as far as values are available. Type 4b: R&D lagging behind, business oriented. No sizable public research activities. As to be expected, differences in R&D intensities at the NUTS 2 level vary widely: between 8,7% in Braunschweig and 0, 1% in Swietokrzyski. Based on the NUTS 2 classification, which is used for the typology clustering , the regions with the highest R&D intensity( GERD in % of GDP) are -with the exception of Berlin - not the large metropolitan cities, but regions which are dominated by a city (like Munich in Oberbayern, or Uppsala in South Sweden ), or cites as such like Stuttgart or regions like Braunschweig, dominated by one big company (VW). BERD dominates in the best performing regions. 8 of the 10 best performing NUTS 2 regions in respect to GERD in % of GDP figure also under the TOP 10 regions in respect to BERD in% of GDP, Among the regions with the highest Public R&D intensity (Government and Higher Education), only 4 are listed under the overall 10 TOP performers. South Europe: astonishingly many Type 4 regions, Portugal: 4, Italy: 13 (out of 21), Spain: 13 (out of 17). 16 Typology of European NUTS 2 regions, Average14 Values by Type Publications per Million inhabitants (2003) Patents per Million inhabitants (2001) GERD as % of GDP (2003, UK 1999) BERD as % of GERD (2003, UK 1999) Outliers n=6 1497.8 579.9 5.24 83.52 Type 1 Region n=25 1843.4 323.0 3.23 67.49 1a n=19 1616.9 284.4 2.98 66.64 1b n=6 2560.6 445.3 4.00 70.18 Type 2 Region n=16 2393.4 114.4 1.64 37.14 Type 3 Region n=59 806.2 121.1 1.40 62.17 3a n=37 916.6 88.7 1.22 52.21 3b n=22 620.6 175.7 1.68 78.94 Type 4 Region n=84 379.3 32.3 0.66 45.76 4a n=54 465.9 17.3 0.62 28.51 4b n=30 223.5 59.2 0.72 76.82 Total (including Outliers) n=190 909.5 122.3 1.45 54.19 Total (excluding Outliers) n=184 890.3 107.4 1.33 53.23 Shadings: dark green: > 100% above average, light green: 25-100% above average dark red: >50% below average, light red: 25-50% below average. Source: Own compilation from the Regional Key Figures, 1-2008, European regional research system: Current trends and structures; H. Kroll, Th. Stahlecker; Fraunhofer;ISI 5.3.3. Summary Regional differences in R&D intensities in Europe are mainly influenced by BERD expenditure. Only in a few of the TOP R&D performing regions public sector expenditure is high enough to influence the overall GERD significantly (for example: Berlin). 14 Unweighted means. 17 There are only a minority of regions which can be classified as R&D driven (27 out of 190) and only in a few of them is the public sector expenditure an important source. Scientific specialization seems to be more influenced by public expenditure (influence of large regional universities) than other indicators. Regions lagging behind have a lack of academic institutions and of higher education which diminishes also their ability to absorb knowledge from abroad and to improve innovation significantly. This is true for most of the Type 4 regions. At the country level, the figures suggest that the Nordic countries are those which have the highest specialization in high tech industries. For example Finland shows a positive technological specification in two high tech industries, namely Telecommunication and Measurement Control, and a negative specification in 8 (low and mid tech) industries. Denmark shows a positive specialization in Pharmaceutical Industry and Consumer Industries, Sweden in Medical Equipment and Machinery Tools, while the Netherlands is strong in Audio-visual Electronics, Telecommunications, Optics. In the contrary Germany seems to be very strong in the medium-tech industry15. Cities with large u5 niversities or other excellent R&D infrastructure seem to have an advantage in attracting R&D personnel as well as R&D funds. (“Hot spot”: Berlin). Not surprising that a number of metropolitan areas are quite successful (Paris, London, Berlin, Brussels). This raises the question how much agglomeration counts for R&D; agglomeration does not necessarily mean concentration to one field of science or to one specific industry but can also be seen as a horizontal approach ( “Related varieties” : many of the high tech clusters focus on several research fields; for example the Helsinki region is focusing on nine technology fields16. 6. Technological-led specialization as a wider E.U. policy issue At both the E.U. and the national level the question of specialization policy cannot be answered generally17. The E.U.-27 consists of divergent national and regional situations, which very often make necessary differentiated policy answers even for apparently similar problems. In addition to these internal disparities the E.U. is disadvantaged because of its laggard position in a range of high-tech/core technology areas vis-à-vis world leaders, with adverse consequences in terms of productivity, growth, market positions and business strengths. One of the visible weaknesses of the E.U. is its difficulty to be path-creating or early entrant in a range of new technologies, to make more efficient use of some of these technologies, to achieve similar productivity results, to diffuse sufficiently such new technologies inside its space and to create a culture of technology and knowledge widely spread in the member states. Both these types of divergence are at the origin of many policy concerns at both the E.U. and the national level, such as: 15 Attention should be drawn to the fact that technological specification as calculated by ISI strongly reflects the influence of the industrial structure; further the application for patents is not independent from the specific industry. 16 “Forschungsstandort Österreich”, Comparison of high tech. regions. 17 A question of different nature concerns the distinction of national and EU competences in the area of RIT policies. The analysis here is focusing on the normative aspects of policy, making abstraction of existing institutional limitations. 20 6.2. The risks from neutral policies Specialization policies either for strengthening existing industrial structures or for catching- up to technological leaders, but in particular for promoting emerging technological fields, are associated with risks of failure. This is half of the reality. The other half is that in absence of any action, risks are equally present. Inaction and/or neutrality can also have risks and costs. The difference to pro-active policies is of political, not economic, character. In cases of inaction failure cannot be directly attributed to no-policy. Hence, it takes not the form of a visible political cost, although it can have a high economic or social cost. To the extent that the speed with which the support of appropriate technologies can contribute to answering the global issues is of high importance, the rapidity of policy-making has a social value. In this case, what matters, is not only to avoid repeating the experience with ICT or other technologies, but to produce timely research results meeting extremely important global needs. 6.3. The risks from targeting specialization For many authors, the capability to organize and to manage technological change and innovation is inherent to the above policy questions. Baumol19 writes that innovation is “the recognition of opportunities for profitable change and the pursuit of those opportunities all the way through to their adoption in practice”, while Drucker20 argues that “innovation ..is the only way to convert change into opportunities. This requires that innovation itself be organized as a systematic activity”. While this type of management concerns mainly the business sector, public policies cannot escape the necessity to manage successfully technological and specialization activities. Often the policy rationale is structured around the concept of market failures and/or externalities. The following specific aspects need to be considered: a) The sectoral nature of many types of technical change and innovations, b) The objective to facilitate new technology paths and/or to replace repeated specializations by a specialization breakthrough, which can create significant competitive advantages, c) The importance to enter early into a new technological area, d) The timely development of capabilities in the aim not to be outstripped by competitors, e) The acceleration of emerging evolutionary processes, f) The enhancement of appropriate institutions, infrastructures, interactions, networks, g) The presence of spillovers when an invention/innovation has a higher social than market (individual) value, h) The public good aspects of certain R&D activities. 19 W.J. Baumol (2002), The free-market innovation machine: analyzing the growth miracle of capitalism (Princeton University Press). 20 P. Drucker (1994), Innovation and entrepreneurship (New York, 1994). 21 In all such cases the goals are unlikely to be successfully achieved without directed public actions. Private returns from investing in non-traditional activities can be highly uncertain, private response can be reluctant, and hence, some intervention could be appropriate. The question is how policies can be effective in attending their targets. This is a general question on government efficiency. However, the question on the ‘how’ –the effectiveness dimension- of a specialization policy is also closely interconnected to the question of ‘what’. What should specialization policies envisage? On the one side, an effective ‘how’ can be defined and implemented only if the ‘what’ is also clearly defined. On the other side, meaningful decisions on the ‘what’ cannot be disconnected from ‘how’ policy will be shaped. Policies in favour of ICT, energy, nanotechnology, climatic changes, health are always determined by the specificities of the objective and the capabilities of the innovation system to perform. Basically, the main question is again how to be selective, what risks have to be taken and how efficient policies can be designed and implemented. In fact, both the ‘what” and the ‘how’ point to the need for public policies to provide efficient or smart support to activities but also to anticipate for timely and efficient discipline in order to avoid ‘picking the losers’. The fact remains, that pro-active policies, even if based on market-evidence, are subject to similar risks of failure as horizontal ones. Besides successes, the experience is full of failed attempts to create Silicon Valleys, Third Italies or Science Parks. Apart from serendipity, there are two main factors causing success or failure. The first is the well known problem of public failures. The second is the inherent high uncertainty and unpredictability of foreseeing ‘what a country will be good at producing’ or ‘where the advantages can lie’. Rarely could successful specializations be visible ex ante. Even countries with very similar levels of technological capabilities and factor endowments followed very different specialization patterns as a consequence of different historical evolutions, entrepreneurial initiatives and policy responses. 6.4. An horizontal issue: Concentration of resources versus decentralization approaches The policy dilemma whether with regard to technology and/or geographical areas, resources should be concentrated around specific goals or be allocated in a more decentralized way, is continuously debated. However, the question is of a more general nature and is not specifically linked to technology policy. The same question has been the object of many other policy debates and is linked to many other concepts (such as indivisibility, critical mass, externalities or rational allocation of resources). It encompassed several theoretical approaches (e.g. the infant industry argument, the industrial pole theory, different other strands of the development theory) and, also, it has been an objective of various policy experiences. Nevertheless, a concentration of efforts to enhance technological development, specific activities, sectors or regional centers is a very common policy practice, both at the national and the E.U. level. The issue in all these approaches is the degree of relative concentration and the mix between focused and decentralized support mechanisms. The policy question is whether we can specify conditions under which the one or the other policy mix should be privileged. Besides, such choices are subject- and time-dependent, and they don’t prejudice choices on other policy issues, or even on the same issue in different time. Hence, 22 concentration or decentralisation policies can coexist and be appropriate, depending on the specific environment. Concentration could be an appropriate tool for situations, in which the failure risks are relatively low and/or the expected benefits have a high social value. In contrast, decentralized approaches can be more relevant when unpredictability and uncertainties are high, or “let all the flowers bloom” policies are appropriate. The concept of ‘smart specialization’ is right in its emphasis on the importance for countries not to imitate, not to duplicate or to reproduce specialization patterns which have been developed in many other countries and regions. However, one has also to take into consideration that what appears as a wasteful duplication can also generate competences which bring a country and/or firm beyond the stage of simple duplication and gives rise to unpredictable innovative capabilities and specializations. In fact, although the vast majority of actors are technology followers or laggards, many of them succeed to build own new capabilities and competitive advantages around ICT, biotech, mechanical, chemical or other areas. What is crucial is that technology change is an open ended process related to a continuous expansion of the frontiers of knowledge, further changes and innovations, even in areas in which already many other actors lead. Hence, policy choices restricting followers for entering dominant areas of technology in the aim to avoid a possible duplication of actions and a waste of resources can be risky. More generally, the view of specialization as a dynamic process indicates that in order to move from a weaker to a stronger (more competitive) specialization position and to ensure long-term advantages, what matters is the capacity to transform patterns of specialization, the capability to shift from existing to new specialization areas and the related more general concept of the capability to change. Path-dependencies can facilitate such a shift and provide a relatively good basis for policy decisions. However, while path dependency can be an important explanation for specialization and competitiveness patterns, it cannot sufficiently explain or guide the development of new core technologies. Neither can it guide catching-up processes of technologically weak countries which are not characterised by previous strong path dependencies. The weak path dependencies in these countries and the configuration of their technology and innovation subsystems indicate that the potential spectrum of new successful specializations is not easily predictable. The policy implication is that risks cut both ways and characterise both types of policies: policies targeting ambitious projects for which capabilities and supportive conditions are insufficient as well as policies which don’t facilitate specialization along specific new technologies, precisely in the belief that the external conditions make such a strategy risky. What is central in this debate is that the two policy choices (concentration vs. decentralisation) are a theoretical distinction and that in reality the two approaches are not substitutes but complementary. Policy should remain flexible enough to deal with different realities and targets. The implication of such an approach means that the following options should remain open: a) For technologically advanced countries, regions, areas and/or clusters: A concentration of resources in the aim to accelerate or support projects with relatively low probabilities of failure can complement horizontal policies, widening the opportunities for a broader spectrum of technological activities. 25 specifically predefined content with regard to sectoral specialization, to areas of gravity or research objectives. The lack of an explicit specification of such priorities is not to confuse with a vacuum. The areas of gravity are implicitly outlined. Since the objective is to make of the E.U. ‘the most dynamic and competitive knowledge-based economy in the world capable of sustainable economic growth with more and better jobs and greater social cohesion..’, this by definition implies the need for the E.U. to develop capabilities on those (existing and/or emerging) scientific and technological trajectories, the dynamics of which drives forward economic growth and welfare in the present era. In this sense, the various high tech (and, selectively, even medium to high tech) areas occupy a central place in the implementation of the Lisbon and ERA goals. In fact, various thematic areas and other initiatives constitute significant priorities of the Framework Program or of the E.U.’s broader research and technology policy. In other words, in order to meet the ERA and the revised Lisbon goals, research, technology and innovation policies have to target both the strengthening of existing technology and industrial structures (depending on their maturity, perspectives and other features) and the facilitation of deeper technical changes. Experience has showed that being a latecomer in core technologies can have serious implications which last for long, are difficult to reverse and affect economic and social performance. In the example of ICT technologies it became apparent, that followers don’t succeed in avoiding productivity and competitiveness gaps as against leaders. Foremost, technological leaders, because of their leadership, are facilitated to expand into new science and technology fields and create conditions for reiterating such processes in further emerging science and technology areas. Basically, many frontier technologies are continuously developed and transformed. Depending on the size and the capabilities, who will gain early enough strong positions in the evolving technological and production areas will be also capable to draw significant benefits both of economic and non-economic nature. Obviously, in the case of emerging technologies market signals are not yet strong to guide policy-making and risks of failure are higher. On the other hand, core technologies take many decades to be developed and the history of technological evolutions shows that: - Rarely, if ever, core technologies have been developed without sufficient public supportive mechanisms, - During these long periods multiple interactions between technological change and public policies are to be seen, which cannot be classified within one discernible pattern, - Often public policies alternate between pro-active and reactive forms according to the specific evolutions and needs, - In this interaction, even reactive policies can under certain conditions have also a pro-active nature. The risks: Technology and innovation policies along these directions imply different risks. Policies aiming at frontier technologies because of weak path-dependencies face increased risks. The high uncertainties for private actors in such situations can make intervention appropriate, but not necessarily less risky. The options: To deal with such risks, policy could be structured along three broad axes: 26 a) To target ‘winning situations’, by leveraging the success of clusters consisting of market players in particular technological, knowledge and specialization areas, based on market-led pre-selection, the evolving market evidence and in cooperation to market agents. What matters is to spark and to underpin a self-sustained cumulative development of new specializations and/or to assure a critical mass of capable market agents. A policy linking selectivity to targeting is very different from policies aiming at ‘picking the winners’. The more significant differences are that23: - Targeting is based on market evidence, on market-led pre-selection and existence of capable market forces and aims in leveraging the success of market players in particular areas, - Government interventions in crucial transition points can have a significant impact on the effectiveness of technology and innovation activities, - Targeting takes place in cooperation to market agents, - It may be important to assure a critical mass of capable market agents. In view of these aspects, policy can favor the enhancement of broader ‘winning situations’, whereby the target is not to identify with individual firms or products, but with technological activities, technological areas or clusters of knowledge and specialization. b) To broaden the policy spectrum by “evolutionary targeting”, in the sense e.g. to assure a critical mass of capable market agents, to target the emergence or to leverage the success of new multiagent structures (or clusters) in particular areas. Evolutionary targeting “operates by enhancing market-led variety and pre-selection through horizontal policies, and accelerating market-led selection and development processes through coordination activities, targeted incentives, institutional changes and other policies”24. c) To combine a) and b) with smart policy initiatives and specializations25. ‘Smart specialization’ indicates a successful fine-tuning of policies envisaging the creation of competitive productive units, clusters and/or regions. It implies some kind of intervention, and, hence, some explicit or implicit targets coupled to some form of intended concentration of resources. The concept of smart specialization (and in general smart policies) is attractive but has various practical difficulties. It assumes that we have the criteria to judge which specialization is smart and which not, and, consequently, which targets are smart. Smart policies can be acknowledged as such only after their success becomes visible, that is ex post. It is very difficult to forecast the Darwinism of the markets during the stages from research to innovation or the combined outcome of market and policy processes. Policies can fail in a short period but be successful, and, hence, smart, over a longer period. Policies might fail 23 G. Avnimelech, M. Teubal (2008), Evolutionary targeting, J. Evolutionary Economics, p.160ff. 24 Ibid., p. 152ff., 160ff. 25 D. Foray, ‘Les nouveaux centres mondiaux dans le domaine de la recherche et de l’ innovation: vers une economie de la specialisation intelligente (FutuRIS, 2008)’ and ‘Understanding “smart specialization” (July 2008)’. 27 because of bad implementation, although specific specializations might be smart and achievable through more efficient policies. Has smart to be defined in terms of technology excellence, of market success, of follower’s strategy? On the other hand, in fact, there is an accumulated knowledge on the characteristics of smart and successful specialization policies. In reality, policy design has to strike a balance between the extremes: political voluntarism and agnostic approaches. The question is how to define the specifications of the needed ‘smart specializations’ in the framework of a particularly blurred research and technology landscape. 6.5.2. Preventive strategies to face global risks: In this phase societies are faced with the need to develop technologies for dealing with qualitatively new global risks (climate change, energy, environmental issues), which enter more and more in the world agenda26. The recent crisis accelerated this process. What is at stake today for leading actors differs from the race to create new knowledge as an engine for growth. It transcends the issue of national or European convergence to the USA or the enhancement of technological capabilities and new knowledge for achieving higher standards of living. In the past, the leaders in new core technologies could achieve higher standards of living, create conditions of a more dynamic evolution, draw benefits in terms of political power, influence and welfare, and, depending on the policy values, enhance the social state. In the present phase the development of technological capabilities to face the global risks associated to the climatic change, environment and energy, and their broader implications on health, food, water and other aspects (massive migration, conflicts) is a qualitative new issue. The risks associated to slow progress on what can be regarded as critical technologies are of a more generic nature, implying a potential deterioration of economic and social situations nationwide, E.U.-wide and even worldwide. Some particular technological fields are today of such an important priority that a failure has high social costs at all levels (E.U., national, regional eventually also global). Briefly, because of the high social priority of some knowledge areas, pro-active (targeted) R&D specialization policies in these fields are not simply an economic, but also a significant social priority. In such cases we are faced with a kind of public goods, which require a policy response dealing with externalities, which the market alone is not able to offer. The key issue is the need for solutions within predetermined time limits, if social costs have to be kept within an acceptable range. A difficulty is that in the case of expected global risks it is inherently difficult to have an ex ante measure of what is success or failure. In the case of new technologies able to meet global risks the measurement of potential benefits and costs differs profoundly from the 26 “European research policy should be deeply rooted in European society. Besides the pursuit of scientific excellence, European research should support knowledge advancement and dissemination and underpin policies for sustainable development in fields of major public concern such as health, energy and climate change” (ERA Green Paper). 30 to compete at the level of scientific and very basic technological discoveries. The other half is that the building of capabilities in new technological areas is not to be regarded in relation only to the early phases of radical and/or significant technical changes. The early phases at a later stage are followed by incremental technical changes, different applications, innovative combinations of existing knowledge or the generation of new specialization niches within the dominant technology fields. In all such examples it is quite rational for followers to create capabilities related to these technologies. As technological change evolves, capabilities of the followers become a moving target, since they have to be continuously improved even for keeping their relative position stable. Notwithstanding such successful examples, horizontal policies appear, however, to be a less risky approach for technologically weaker systems. They generate decentralized selection mechanisms, learning processes and a diversification of specialization patterns, while they also facilitate innovative forms of combinations between technological knowledge and local factor capabilities. 7. A lesson from the financial crisis: The importance of regulating risky innovation-based specializations The present financial crisis, and particularly the mechanisms underlying its origins made more explicit an additional policy dimension with regard to technological specialization policies. They showed that besides the question of ‘what policies for good specialization, innovation-based specializations have to be judged not only on their growth-generating potential, but also on their risks. This implies the need to complement short-term criteria for policy-making with medium- or long-term considerations. At the origin of this crisis one can find the significant role of financial innovations, as expressed in the new financial products, including the derivatives or the securitization of loans and the spread of these titles and their associated risk worldwide. Governmental practices followed the typical device to enhance promising innovations: they proceeded to significant institutional changes facilitating the expansion of these products, even more since they ensured a particular impetus to growth. New, more lax, regulations on supervision, operational functions, financial leverage practices were introduced, accepted and adjusted to the perceived new growth opportunities. As known, the result was the creation of an innovation-induced growth mechanism, which after reaching its limits, caused an unprecedented recession, bankruptcies and unemployment, with significant destabilizing impact on large parts of the world economy. Taking into account the fact that the most serious destabilizing phenomena during the last twenty years (starting with the nationally based Tequilla crisis in 1995) were of financial nature, a lesson to be drawn is that what matters is not only where and in what to specialize or how the creation of temporarily smart ‘differences’ can enhance innovative specialization patterns. In addition, the issue is how governments can frame innovative specializations and control their applications, in the aim to prevent serious negative effects (social, economic, environmental or other)28. More precisely, the issue is that policies on good specialization 28 A different, albeit related, example is the institutional regulations introduced in the case of genetic cloning. 31 should comprise the capability to perceive, prevent and protect from bad or risky specializations. 8. What choices and how to deal with risks? Frontier research is not a question of percentage spending to GDP but of smart goals and policies as well as of appropriate absolute amounts of financial and human resources. Evidence shows that voluntaristic top-down approaches have often failed, but that neutral policies often have also a failure cost, except that this takes less transparent forms. The success of both, target-related and neutral strategies depends largely on the articulation of the policy mix and the definition of the objectives. Faced with different asymmetries of information, risks and opportunities, policy-making can be addressed as a risk management issue and could draw from the concept of ‘portfolio management’, adjusted to technology and innovation policies. Portfolio management approaches favor variety and selection mechanisms, can reduce risks and assess the multiple research and technology objectives on the basis of such criteria as financial cost, probabilities of success, externalities and/or social costs and benefits. An interesting distinction in this respect is between related and unrelated varieties, because it reflects the very diverse realities between technological leaders and followers and the factors determining the rationality of different kinds of policy29. Even if technology and innovation activities by their nature are high risk activities, firms are expected to follow Schumpeterian behavior and intensify research and innovation. In a Schumpeterian-driven economy the State cannot ignore the need to select specific goals and means and to take also its risks in this process. The selection by the market only may fail by reasons like externalities, public good characteristics etc. and the selection by policies may fail as well. Hence, for public technology and innovation programs a portfolio management may provide a useful instrument through which risks will be spread and minimized, especially when a close cooperation between politics and the business sector is secured. The main question still is how to shape targets and choices, which can better reflect the politically accepted balance of policies, social risks and benefits. In view of the mentioned three major E.U. challenges the question is if and what new policy concepts have to enrich or to enhance the existing policy-making process, how policy could better succeed in organizing a flexible and diversified framework and implementing specialisation targets30. Our analysis implies also that in the framework of the ERA the support of high tech clusters is a necessary but not sufficient condition for successful policies. What is essential is to 29 K. Frenken, F. van Oort, T. Verburg (2007), Related variety, unrelated variety and regional economic growth, Regional Studies, 41, 5, 685-697. 30 Pro-active policies at the EU (and national) level can aim at a ‘research friendly ecology’ (Georghiou, 2007), and be selectively combined with a ‘cluster-specific environment’. 32 shape governance structures which can conceive timely and effectively the necessary R&D and technological policies31. Success is codetermined by a range of additional elements: • An appropriate coordination at European level of public organizations, business firms and research communities, each of which has different interests, priorities, or strategies (e.g. on the appropriation of new knowledge) and functions, • The design of priorities on selected areas and a package of policies to support the research activities of firms and organizations and a close cooperation with the business sector and the scientific community in detecting needs, capabilities, technological trends, key discoveries, possible advancements, • Policies promoting existing or emerging technologies, instead to proceed on a voluntaristic base have to rely on the signals of the (research, innovation, product) market, • For the evaluation of success, the selection of priorities as well as other policy strategies has to consider externalities -positive as well as negative ones-, like climate change, energy supply, environment issues. Within the concept of portfolio management, the effects of these externalities have explicitly to be taken into account, • The broadening of the criteria on the basis of which the success of research and technology specialization policies can be assessed, especially with regard to the coherence, the efficiency, the long-term commitment of these policies as well as the time framework within which policies have to be implemented, • The enhancement of variety creation and the selection and support of differentiation elements vis-a-vis competitors, and • The capability to design and implement effectively appropriate policies and governance of these policies (i.e. the national innovative capacity) in a long-term period, since the effectiveness of policies often depends on: a) the way they can meet successfully the above conditions, b) the complementarities with other policies, c) the broader economic environment, d) the supportive activities, and e) the social capabilities to adjust and to exploit opportunities. The problems are different for big and smaller countries. Big countries benefit from the advantage to have more large areas where research can be distributed. In contrast, smaller countries have fewer choices to differentiate their technology and innovation activities. Consequently, they are characterized by a more concentrated and more risky pattern of technology and innovation choices. Big countries can follow a differentiated policy with regard to their different regions. Even a regional polarization -in terms of research, knowledge, clusters and/or sectors- has different implications for bigger and smaller 31 See for example the principles of ‘governance and trust’ proposed by R. Marimon, M. de Grace Carvalho (2008), p. 9ff. 35 Godor et al., (2005), Grundlagenforschung in Österreich (März 2005). Haque I., (2007), Rethinking Industrial Policy (UNCTAD, Discussion Paper No.183). Hausmann R., Rodrik D., (2002), Economic development as self-discover (NBER, WP 8952). Kaiser R., Prange H., (2004), The reconfiguration of National Innovation Systems –the example of German biotechnology, Research Policy, 33., 395-408. Kroll H., Stahlecker Th., (2008), Regional Key Figures 1-2008 (ISI Fraunhofer) Kuhlmann St., (2001), Future governance of innovation policy in Europe –three scenarios, Research Policy, 30., 953-976. Lakoff S., (2008), Upstart startup: “Constructed advantage” and the example of Qualcomm, Technovation, 28., 831-837. Larosse J., Pontkakis D., R&D specialization: strategic intelligence in priority setting (OECD TIP-RHIR Workshop, Sept. 2008) Leo H., Reinstaller A., Unterlass F., (2007), Why does sectoral analysis of innovation performance matter? (WIFO). Malerba F., (2005), Sectoral systems of innovation: A framework for linking innovation to the knowledge base, structure and dynamics of sectors, Economics of Innovation and New Technology, 14., 1-2, 63-82. Malerba F., Montobbio F., (2000), Sectoral systems and international technological and trade specialisation, (Paper, June). Marimon R., Graca Carvalho M., (2008), Governance and coordination of S&T policies in the European Research Area (Paper, KfG Group, June). Nelson R.R., (2008), What enables rapid economic progress: What are the needed institutions? Research Policy, 37., 1-11. Nelson R.R., Pack H., (1998), The Asian miracle and modern growth theory (The World Bank, PRWP 1881). Pack H., Saggi K., (2006), The case for industrial policy: a critical survey, (World Bank, PRWP, 3839). Paoong Chang, Cienz-Tzu-Tai, (2002), Funding a Niche Position-Competitive Strategy of Taiwan`s IC Design Industry”, Technovation, 22., 101-111. Peter, V., Frietsch R. 2008),”Exploring regional technology specification. Implications for policy (Fraunhofer ISI). Pontikakis D., Chorafakis G., Kyriakou D., (2008), R&D specialization in the EU: from stylized observations to evidence-based policy (IPTS Discussion Paper). D. Rodrik, (2004), Industrial policy for the twenty-first century (UNIDO). Storz C., (2008), Dynamics in innovation systems: Evidence from Japan’s game software industry, Research Policy, 37., 1480-1491. Zucker L.G., Darby M.R., Furner J., Liu R.C., Ma H., (2007), Minerva unbound: Knowledge stocks, knowledge flows and new knowledge production, Research Policy, 36., 850-863. 36