1. SECTORAL SYSTEMS AND CATCH UP Franco Malerba Globelics Academy, Tampere, June 2015

2. THIS LECTURE Do sectoral differences matter? Building block 1: technological regimes Building block 2: sectoral systems Sectoral systems and catch up Sectoral systems and successive changes in industrial leadership Windows and responses The four challenges ahead Sectoral systems Catch up in different sectoral systems

3. Catching up, lagging behind and forging ahead The meso and micro view: the growth of local knowledge and the changing productive and innovative structure of the economy. Catching up does not mean imitation or cloning. Catching up means a process through which emerging countries learn and accumulate knowledge, and develop products, processes and technologies that may differ more or less from the ones of advanced countries. The focus on learning and knowledge implies the possibility of different trajectories of knowledge accumulation, innovation and production specialization with respect to the leading countries. ECONOMIC DEVELOPMENT AND CATCHING UP

4. TO DAY FOCUS ON MESO AND MICRO The focus on meso and micro implies that development and catching up: - may take place in industries that can be different from country to country and that may change in importance over time - means that the learning processes and the factors that affect the catching up may greatly differ from one industry to another

5. Why sectoral innovation systems? 1.Innovation in sectors is the result of different learning processes and capabilities, the use of different knowledge and different actors and their interaction 2. Major differences across sectors exist. Think for example of ICT, pharmaceuticals or textiles What do we know about these issues ?

6. Various traditions have emphasised sectoral differences Case studies Innovation studies –Keith Pavitt Industrial economics Management literature and Porter’s five forces

7. WE WILL MOVE FORWARD In two in steps. First, we will consider the learning and knowledge environment of firms in sectors. These learning environment may differ across sectors Second, we will consider non only the learning and knowledge environment, but also industries as systems

8. The learning and knowledge environment that affects innovation This environment may greatly differ across sectors

9. TECHNOLOGICAL REGIMES Opportunities (scientific and technological) - Level - Sources (R&D, universities, research institutes,…) Appropriability of innovations - Level - Instruments (patents, secrecy, …) Cumulativeness of technical advance Cognitive base underpinning innovative activity in a sector Nature of knowledge: some simple distinction - Complex / Simple - Tacit / Codified - Independent / Part of a system

10. FROM CONCEPTS AND THEORIZING TO MEASUREMENTS AND ECONOMETRICS F. Malerba L.Orsenigo Industrial and corporate change 1997 S.Breschi F.Malerba L.Orsenigo Economic Journal 2000 Castellacci F. Technological regimes and sectoral differences in productivity growth Industrial and Corporate Change 2007

11. THE RELATIONSHIP BETWEEN TECHNOLOGICAL REGIMES AND SCHUMPETERIAN PATTERNS BASIC RELATIONS from Malerba and Orsenigo (1997) and from Breschi, Malerba and Orsenigo (2000 ) CONC= f (OPP,APP,CUM,Kbasic, Kapplied) STAB = f (APP, CUM) ENTRY = f (OPP, CUM, KBA, KAP) ? + ++- ++ +-+ -

13. TECHNOLOGICAL REGIMES AND DIFFERENCES IN PRODUCTIVITY GROWTH Technological regimes (APP, OPP, CUM) External sources of opportunities (cooperation, users, science and suppliers) Industry specific characteristics Skills and training Openess Market size Country specific factors Castellacci, ICC 2007

15. Differences in innovation and production across sectors: sectoral systems

16. F. Malerba Sectoral systems of innovation and production Research Policy 2002 F. Malerba How innovation differs across sectors: sectoral systems of innovation Handbook of Innovation Fagerberg, Mowery, Nelson (ed) Oxford University Press 2005 F. Malerba and P. Adams Sectoral systems of innovation magement in Oxford Handbook of Innovation Management Oxford University Press 2013

17. The foundations Change and transformation in industries Industry life cycle Innovation systems Lundvall, Edquist Evolutionary theory Nelson and Winter, Dosi

18. SECTORAL SYSTEMS OF INNOVATION AND PRODUCTION Innovation and production take place in sectoral systems that are characterized by: -a knowledge base -firms with learning and capabilities -other actors relevant for generating or supporting learning and innovation in a sector such as suppliers, users, universities, public research organizations, the government, financial organizations; -links and networks among the different actors; -institutions; - processes of interaction, cooperation, and selection Major differences exist among sectoral systems

19. The central role of knowledge The specific knowledge base of sectors The content of knowledge The sources of knowledge Specificity, tacitness, complexity, complementarity, independence Knowledge and the boundaries of sectors

20. Dynamics and evolution: key aspects of sectoral systems Sectoral systems: dynamics first ! Dynamics and the role of: - knowledge, learning and capabilites - various actors - demand The redefinition of the boundaries of sectors

21. A sectoral system approach for the study of innovation in sectors: -Is not a straight jacket, but broad, flexible and adaptable: it points to a system view of sectors, to key variables and to fundamental relationships -It leads to analyses done at different levels of aggregation, depending on the purpose of the research -It has a strong link with theory -It allows modeling -It enables quantitative analyses -It allows for comparative analysis across industries and across countries -It provides a framework for policy -

22. WE KNOW A LOT OF SECTORAL SYSTEMS OF INNOVATION IN THE UNITED STATES AND EUROPE F. Malerba Sectoral systems of innovation Cambridge University Press 2004 Wide variety of papers on specific sectors

23. AND SOMETHING ABOUT THE RELATIONSHIP BETWEEN SECTORAL SYSTEMS AND ECONOMIC DEVELOPMENT F.Malerba and S.Mani “The structure and evolution of sectoral systems in developing countries” Elgar 2009 Pharmaceuticals, telecommunications, ICT, software, aeronautics, motorcycle, salmon farming, machine tools, bio-fuels Various countries examined.

24. Major insights from Malerba and Mani (2009) Understanding the specificities of the relevant sectoral system is fundamental in order to identify the source of innovation in development The so-called traditional sectors may have a process of major transformation and be highly innovative The separation of research from development and production capability can be very harmful for innovation and development The type of network that emerge in a sector is strongly associated with the specific knowledge base of that sector Vertical interdependencies and local demand may foster in some cases or impair the full development of a sector

25. Sectoral systems and catch up Malerba and Nelson: Catching up in different sectoral systems Industrial and corporate change, 2011 Malerba and Nelson: Economic development as a learning process Elgar 2012

26. Pharmaceuticals: India and Brazil S.Ramani and S.Guennif Agro-food: China, Brazil, Nigeria and Costa Rica S.Gu, J.Adeoti, A.Castro, J.Orozco Software: India, Brazil, Russia, Philippines... G.Niosi, S.Athreye, T.Tschang Telecommunications: Korea, China, Brazil K.Lee, S.Mani, Q.Mu Auto: Korea, China, Brazil R. Quadros, Y. Huyn, Y. Wang Semiconductors: Korea, China, Malaysia R. Rasiah, X. Kong, Y.Lin

27. COMMON FACTORS ACROSS SECTORS AFFECTING POSITIVELY CATCH UP 1. Accumulation of domestic capabilities 2. Access to foreign knowledge and international links 3.Development of advanced human capital 4.Active government policy In successful countries, these factors work together in a dynamic complementary way: learning and capability building, knowledge access, advanced human capital

28. SECTOR-SPECIFIC FACTORS (1) a)INDUSTRIAL STRUCTURE - Scale and large size vs entry of new small firms - Presence of local clusters - Role of multinational corporations (global value chains vs licences to local firms vs competent local branches) These differences are the result of the working of -different technological regimes, of different scale, scope and modularity of production and of different demand -dynamic elements such as industry life cycle and the time of catch-up

29. b)DEMAND CONDITIONS: -Different role of exports, domestic demand, demand segmentation and user-producer links. -Often a dynamic relationship between demand and economic performance has been present c)OTHER SECTORAL SYSTEM ELEMENTS - Universities and public research organizations (pharma and telecom) - Finance: venture capital vs internal finance (software vs telecom and auto) - Diversity in the types of government policy (relevant in telecom vs software vs pharma), with different dynamic effects over the short run and the long run SECTOR SPECIFIC FACTORS (2)

30. AGAIN, THESE FACTORS ACT TOGHETHER IN COMPLEMENTARY WAYS SOFWARE: entrepreneurship, local clusters, links with customers TELECOM: large firms, active government support for domestic firms R&D SEMICOND: entrepreneurship, specialization PHARMA: university research, relaxation of IPR AGRO-FOOD: development of market institutions, diffusion policies, development of a knowledge infrastructure

31. A KEY ISSUE: THE ROLE OF THE NATIONAL INNNOVATION SYSTEMS The role of national systems in terms of national institutions and government policy has been significant across sectors For example: government policy in Korea has supported learning and capability accumulation by large firms, in Taiwan new small firms and in Brazil has been favorable to multinationals This has affected also the sectors of a country catching up. There has been a two-way dynamic relationship between national and sectoral variables: national policies have affected specific sectors that in turn have pushed for specific national policies

32. Sectoral systems and successive changes in industrial leadership Joint work with Keun Lee

33. MOBILE PHONES 33 7.1 9.7 10.5 12.6 12.7 11.8 13.4 16.3 19.5 17.6 17.7 22.0 35.0 35.1 34.7 30.7 32.5 34.8 37.8 38.6 36.4 28.9 23.8 19.1 2.9 5.0 7.5 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 200120022003200420052006200720082009201020112012 (%) Source: Gartner presentation 33

34. STEEL (1) Sources: 1900-79 – Mitchell (1995, pp. 456-62, 1992, pp. 417-19); 1980-2010 – World Steel Association

35. STEEL 2 Sources: 1973-4, POSCO – Song (2002, p. 150); others – World Steel Association

36. In several sectors there have been successive catch-ups and changes in industry leadership. And often companies come from the same country, therefore supported by the same national innovation system Mobile phones: Motorola -> Nokia -> Samsung and Apple Steel: U.S.  Japan  Korea  (partly) China Automobile: Germany  U.S.  Japan  Korea  partly China Shipbuilding: U.S.  Britain  Japan  Korea  partly China Semiconductor (memory chips): US-> Japan -> Korea Franco Malerba

37. Three questions: Why do successive catch-up cycles occur? What are the factors that lead to successive changes in industry leadership in a sector ? Is it possible to model these cycles? And understand the public policies that affect them? Franco Malerba

38. Industry Catch-Up Cycles

39. Factors affecting catching up: emerging views Capabilities: Key role of domestic firms in catching up: learning, capabilities and strategies are fundamental for catch-up L. Kim, K.Lee, National innovation systems: Importance of institutions and the relevance of national innovation systems for catch-up C. Freeman; B.A. Lundvall; R.Nelson But major differences across sectors exist. Evidence from: -Sources of industrial leadership: Mowery and Nelson (1999); -Sectoral systems and economic development: Malerba and Nelson (2012) Franco Malerba

40. Looking for explanations for successive catch-up cycles: our starting point For an explanation, we need a firm, a country and a sector view and consider diverse factors A starting point is the concept of Sectoral Systems of Innovation (Malerba, 2002), that has firms as central actors and system factors affecting firms innovation and performance. Building blocks of sectoral systems: –Firms: learning, capabilities, strategies –Other actors: universities, research centers, users and consumers, suppliers, financial organizations, government agencies… –Knowledge base and technologies –Institutions (public policy, IPRs, regulation…) –Systemic effects: interactions and interdependencies Franco Malerba

41. Initial idea Leapfrogging and Window of Opportunity (Perez and Soete 1988) “Techno-economic paradigm changes can be windows of opportunity for late-comers. They may bypass the old paradigm and jump into the new paradigm and thereby leapfrog”

42. Three windows of opportunities at the sectoral level for catch-up 1) Radical changes in the knowledge base and technologies a. New Techno-Economic Paradigms (Perez and Soete 1988) b. Generations of new technological trajectories, disruptive innovations Analogue  Digital: Korean Digital TV (Lee, Lim and Song, 2005) 2) Changes in Demand a.Changes in demand conditions Emergence of Chinese market with specific demand b. Business Cycle Downturns TFT-LCD Industry (Mathews, 2005) 3) Changes in government intervention/regulation Examples: Indian pharmaceutical industry (Guennif and Ramani, 2010) - Telecom in China, India, Brazil and Korea (Lee, Mani and Mu, 2011) Franco Malerba

43. Crystal cycles and Late Entries during Downturns

44. Downturns in business cycles as windows of opportunity Mathews, 2005 Downturns provide a time for economic cleansing and entries Brake on incumbents Release of resources provides opportunity for challengers – newcomers and latecomers Technology Transfer and Knowledge - Access become easier and cheaper Downturns provide opportunity for fast followers to create supply chain and to move up in rankings Strategy for latecomers – timing of entry, making use of released resources and knowledge; Example: 2008/9 Global Crisis  window for China

45. Windows are always doomed to open sooner or later but firms and countries may or may not be able to catch them Windows are always doomed to open as new technologies may be introduced, demand may change and government may start playing a major active role in catching up Leadership change and success in catch-up depend on a combination of windows of opportunities with capabilities and strategies of incumbents firms and of catching up firms

46. THE RESPONSES TO A WINDOW BY INCUMBENTS AND LATECOMERS Levels of capabilities System response

47. Learning and Capabilities For Catching up Firms: learning and capabilities are key drivers of catching up This process is gradual and may take a long time It is based on production and R&D investments and on human capital investments See the work of L. Kim (1997), K. Lee and others Learning and the accumulation of capabilities by firms may be affected by system factors (education, public policy of support, diffusion policies….) For incumbent firms Winners’ curse/ trap (not necessarily mistakes): the leaders tend to be complacent and entrenched with current success. They do it also by an obvious choice, given the uncertainty of new technologies or nee markets and given fixed investments whose life cycles have not finished yet Franco Malerba

48. Lee and Lim (2001) Path of the Forerunner: stage A --> stage B --> stage C --> stage D 1. Path-Following Catch-up: stage A --> stage B --> stage C --> stage D e.g. PC, some consumer goods, and machine tools industries in Korea 2. Stage-Skipping Catch-up: stage A -----------------> stage C --> stage D e.g. Hyundai's fuel-injection engine development (cf. carburator engine) Samsung’s 64K DRAM prod. technology; 256K DRAM design technology 3. Path-Creating Catch-up: stage A --> stage B --> stage C' --> stage D' e.g. Korea’s CDMA and digital TV development C and C‘ represent different or competing technologies Also the strategies of catching up firms may prove quite important

49. System factors Level of education Supporting policies Financial system Universities and PRO

50. Industry Catch-Up Cycles In a sector: 3 possible Windows of Opportunity and 3 possible Cycles In a sector, each cycle is that of a leading firm or a collection of firms in a nation. A new cycle replaces an old cycle

51. The empirical evidence of successive catch-up cycles Successive catch up cycles have occurred in several sectors. 8 sectors have been examined by a group of researchers. In all the 8 cases examined, catching up firms have developed advanced capabilities through a continuous learning process and deployed successful strategies. In addition, windows of opportunities opened up and allowed a change in leadership. Franco Malerba

52. Mobile Phones (Claudio Giachetti and Gianluca Marchi) Aircraft (Daniel Vertesy) IT Services (Sunil Mani) Cameras (Hank Kang and Jaeyong Song) Semiconductors (Shin Jang-sup) Steel (Jehoon Ki and Keun Lee) Wine (Roberta Rabellotti and Andrea Morrison) Games (Yuko Ahoyama and Hiro Izushi)

53. Successive changes in industrial leadership: the 8 sectors examined Steel industry: Keun Lee and Jeehoon Ki Technological windows (basic oxigen furnace and continuous casting) allowed a leadership change from US to Japan. A demand window (global steel industry downturn) plus a government window generated a leadership change from Japan to Korea Camera industry: Jaeyong Song and Hoyseok Kang A technological window (rangefinder/portable camera) allowed a leadership change form Germany to Japan. A new technological window in terms of a radically new trajectory (compact system camera) generated a change from Japan to Korea Semiconductors (memory chips) Shin Jang-sup In the various generations of DRAM chips, demand windows (business cycles) plus stage skipping strategies allowed a first leadership change from US to Japan, and then a second change from Japan to Korea

54. IT services Sunil Mani, Aborted catch up from US to Ireland and then sustained catch-up from US/Ireland to India Mobile phones Claudio Giachetti and Gianluca Marchi A technological window (digital phones/GSM) and a regulation window allowed the shift from US (Motorola) to Europe (Nokia). A new technolological window (smart phones) allowed the shift from Europe to US (Apple) and Korea (Samsung) Regional jets Daniel Vertesy Government windows allowed a change in leadership from UK to Netherlands (Fokker), then to Canada (Bombardier) and then to Brazil (Embraer) Franco Malerba

55. Wine Roberta Rabellotti and Andrea Morrison Demand windows and technology windows allowed first the catching up of US and Australia and later on the change in leadership from France to Italy Games Yuko Ahoyama and Hiro Izushi A demand window (market crash) allowed a change in leadership from US to Japan (Nintendo and Sony). Later on a technology window (new ICT and software) and a demand window allowed a leadership change from Japan to US again Franco Malerba

56. A first understanding of successive catch-ups Successive catch-up cycles are a combination of (a) windows of opportunities coupled with (b) responses by latecomers in terms of domestic firms capabilities and strategies and by systems More than one window can be relevant for catching up in a specific stage. Entry by latecomers often takes the form of public-private partnership, SOEs or OEM/sub-contracting, often induced by public industrial policy Gradual catch-up led by cost advantages, associated with lower wage costs and to the adoption of new capital equipment. It tends to be concentrated in low-end segments (path-following strategies) Rapid catch-up is triggered with a shift of technology and demand conditions, combined with strategy of a path-creation or stage-skipping

57. Finding 1: Windows are different Technology windows: Camera: Japan SLR camera replacing German RF; Cell phone: shift from analogue to digital (from Motorola to Nokia) Steel: innovations of BOF and CC: US -> Japan Demand Windows India IT service: Y2K and dot.com boom Korean steel: downturns after oil shocks New World Wines: rise of new inexperienced consumers from the UK, the USA and the Scandinavian countries; then lately rise of Asian markets Government Windows: Korean steel, Brazilian mid-size jets; Nokia (EU standard)

58. Finding 2 : Incumbent Traps are frequent Likelihood of leadership change increases when the technology windows are combined with ‘incumbent/winner trap ‘ behavior In their early days, new technologies are often more costly, less productive, and less reliable. Thus, the incumbents who command the highest productivity from the existing technologies feel no reason to adopt new technologies. They are complacent with the current success and dominant technologies. It is not certain whether the choice by the incumbents is simply a mistake or an ex ante rational decision-making, given uncertainty of new technology or the new market, and given their fixed investments in technologies and products whose life cycle has not finished Examples: Motorola tried to improve further the existing analogue telecommunication technologies despite arrival of digital technologies. Steel: American steel firms did not adopt the BOF; because they constructed many mills with the old method and thus the useful economic life of them did not end yet when the BOF started to be put into commercial operation in the 1950s

59. Finding 3: Also upturn can be a windows? Downturn = window for a stage-skipping entry with lower costs Upturn = window for a leapfrogging because of higher possibility of incumbent lock-in (trap) with existing technologies (during upturn) Then we may have a sequence of catch-up strategy a) path-following entry b) upgrading during downturn for up-to-date technology c) leapfrogging during upturn for emerging technologies

60. Finding 4: Sectoral Specificities of Windows and Leadership Dynamics Sectors differ in terms of the type of windows that most frequently open up and in the type of catch-up cycle. 1)In sectors with technology windows (semiconductors and mobile phones) a high probability of radical replacement of incumbent by new entrants. 2) In other sectors (wine and auto) with demand windows. new firms often co-exist with old incumbents rather than replace them completely

61. Finding 5: Are windows endogenous ? Windows may also endogenously created by actors the current leaders may try to innovate in the direction of competence- enhancing way. -> If the leaders succeed, they are more likely to maintain their leadership in the next generations. -> persistence in leadership Ex 1) Samsung’s memory chip business: leader since the 1992 ; while industry had several leadership changes before the rise of Samsung. Ex 2) Canon’s continued leadership in cameras

62. Finding 6: The Perverse Effects of Macro Conditions Longer term changes in macro-variables can be a factor for both rise and decline of a country. 1)Low wage rates in emerging countries = an initial window for ‘learning by doing’ in FDI/OEM ; but wage rates tend to rise with a country success 2) Exchange rates: under-valued as subsidy to exports. But values of currencies are expected to appreciate eventually if a country succeeds in exporting and thereby records trade surpluses. Ex: 1 Japanese Yen = 2.7 Korean Won in 1980; 5.8 Won in 1988; 7.4 Won in 1996; 10 Won in 2000.

63. The empirical evidence and the appreciative theorizing proposed here centered on: (a)sectoral systems (b)windows of opportunities (c)firms capabilities and strategies will provide a useful inroad to a better understanding of the factors that affect not only catch-up in general, but also: successive catch-up cycles in the same industry and their differences across industries. Franco Malerba

64. IN CONCLUSION For catch up “One size does not fit all”. The same “learning/knowledge/industrial structure/ interaction/ innovation/ institutions/ dynamics” link does not fit all sectors in the same way Diversity is relevant across SECTORS. Major DIFFERENCES exist across sectors in the learning processes, knowledge accumulation, structure, dynamics and transformation.

65. The sectoral dimension mediates between the MICRO analysis of learning and knowledge within firms and by firms and the MACRO relationship between knowledge and economic growth. Sectors affect the MICRO dimension of firms learning and capability accumulation because of sectors’ specific knowledge base, structure and institutional context and the MACRO analysis of knowledge and economic growth in terms of the changing structural base of knowledge accumulation.

66. CHALLENGE FOR RESEARCH - 1 CONTINUE TO ENLARGE THE SECTORS EXAMINED IN NEW LEADING AND DEVELOPING COUNTRIES In particular: Services Traditional or low tech sectors CHALLENGE 2 DEVELOP TAXONOMIES OF SECTORAL SYSTEMS

67. CHALLENGE 3 UNDERSTAND THE DYNAMICS AND EVOLUTION OF SECTORAL SYSTEMS IN THE PROCESS OF ECONOMIC DEVELOPMENT For a sectoral system: is there a unique structure, path, trajectories across countries? Or not? Key themes: understanding the emergence of new sectors, the transformation of existing sectors, the possible trajectories of a sector, the lock-ins, competence destruction within sectors, self reinforcing mechanisms in sectors Examine the specific co-evolutionary processes at the base of change and growth in sectors CHALLENGE 4 PROGRESS ALONG COMPLEMENTARY METHODOLOGICAL TRAJECTORIES THAT FEED ONE WITH THE OTHER 1.Case studies on actors, structure and evolution of sectoral systems 2.Appreciative theorizing 3.Quantitative-econometric analyses 4.Modelling Link 1-2-3-4 together !