SPRU Friday seminar – 22 February 2013

1. SPRU Friday seminar – 22 February 2013 Nicoletta CORROCHER# Grazia CECERE* Cédric GOSSART* Müge ÖZMAN* # KITeS, Bocconi University, Milan * Télécom Ecole de Management, Institut Mines-Télécom, Paris Patterns of innovation in Green ICTs: A patent-based analysis An output of the ECOPATENTS project funded by ADEME http://ecopatents.wp.mines-telecom.fr/

2. Outline • Introduction • Background literature • Methodology & Data • Results • Implications for future research 2

3. Introduction 1. Introduction Context: economic & ecological crises Solution: eco-innovations? (to enable decoupling) Ecoinnovation in the ICT sector: “green ICTs”? Limits of green ICTs: their own ecological impacts Key question: Can innovation dynamics in the green ICT sector contribute to solve the aforementioned crises? Objective of this paper: Analyse dynamics of innovation in green ICTs

4. 1. Introduction ICTs & the environment Source: Hilty (2008: 147). Green ICTs Ecological impacts of ICTs 4

5. 1. Introduction Green ICTs(hardware + software) Reduce negative ecological impacts + create jobs: • Direct impacts: ICTs reduce their own ecological impacts. • Enabling impacts: ICTs are used by other sectors to reduce ecological impacts. • Systemic impacts: The diffusion of ICTs enable structural changes towards sustainable development (absolute decoupling). Source: Hilty, L. (2008), Information Technology and Sustainability: Essays on the Relationships between Information Technology and Sustainable Development, Norderstedt: Books on Demand. 5

6. 2. Literature Ecoinnovation & patents • Increasing number of studies using patents to analyse ecoinnovation (cf. lit. rev. in our paper: http://ssrn.com/abstract=2117831). • The propensity to patent varies across sectors including for eco-technologies (=> which ones are the most dynamic?). • Eco-innovation studies cover different technological domains, e.g. automobile, waste, water, air cleaning (not ICTs). 6

7. Objective & Questions 3. Methodology & Data • Objective: Examine the emergence and dynamics of green ICT technological domains. • Question: What are the patterns of innovative activity in green ICT technological domains, notably in terms of: • Growth of patents • Concentration of innovative activities by organizations and countries • Entry of firms • Technological pervasiveness • Sources of knowledge

8. 3. Methodology & Data Data Construction of our green ICT technological domain: • WIPO Green Inventory– includes IPC codes associated with environmentally friendly technologies (6 technological fields: alternative energy production, transportation, energy conservation, waste management, agriculture/forestry, administrative/regulatory as well as design aspects, and nuclear power generation). • OECD classification of ICTs - includes IPC codes associated with the ICT sector (4 technological fields: telecoms, consumer electronics, computer and office machinery, other ICTs). 8

9. 3. Methodology & Data Methodology 3 steps: • Select EPO patents that have at least one green and one ICT technological class (at 7 digit level) granted between 1987 and 2006. • We perform a network analysis on the classes, where each node is a technological class and each link represents number of patents (we select couples of classes that have at least 55 patents - top 1%) dataset includes, 3978 classes at 7 digits (795 green, 2859 ICT and 325 pure green ICT) and 13210 patents. • We analyse network components (technological domains) in detail.

10. 3. Methodology & Data Methodology • 1. Select EPO patentsthat have at least one green and one ICT • technological class (at 7 digit level) granted between 1987 and 2006

11. 3. Methodology & Data International Patent Classification (IPC) codes ICT IPC codes Green IPC codes a,c

12. 3. Methodology & Data International Patent Classification (IPC) codes Pure Green ICT IPC codes ICT IPC codes Green IPC codes b

13. 3. Methodology & Data 69 67 80 H01S 5/34 Methodology H01L 33/00 80 2. Network analysis: Analyse the structural characteristics of this network. Clusters of codes can define technology domains.

14. 4. Results Network analysis – Detection of Green ICT domains 65 different technological fields (components):

15. 4. Results The components of green ICT technological fields 15

16. 4. Results Patents & IPC classes in green ICT fields 16

17. 4. Results Patterns of Innovative Activity The average number of patents per domain is 412, but the distribution is very skewed (a couple of domains concentrate most of the patents). 17

18. 4. Results Annual growth rate of patents in Green ICTs (1987-2006) Average annual growth rate = 0,34 (34%) 18

19. 4. Results Firms with the highest number of Green ICT patents 19

20. 4. Results Top 3 organisations by Green ICT technological component 20

21. 4. Results Top 3 innovative countries in Green ICT technological domains 21

22. 4. Results Top 3 innovative countries in Green ICT technological domains (cont.) EU 22

23. Patterns of innovative activity in Green ICTs Growth of patents over time Different types of indices: Country concentration (HHI index)* Organization concentration (HHI index)* Entry of new patenting firms Technological pervasiveness: the extent to which domains spread across different IPC classes (Jaffe et al., 1993)* Knowledge source across technological classes-variety of knowledge (Trajetenberg et al., 2002)* Internal knowledge sources Academic sources of knowledge * Using the method suggested by Hall (2002). 4. Results

24. 4. Results Growth: The average annual growth rate of the domain wrt the number of patents HORG: The concentration degree of organizations ENTRY: share of firms patenting for the first time in domain i over the total number patenting HCOUNTRY:The concentration degree of countries in each domain HTECH: the extent to which the domains spread across different technological classes. SELFKNOW: Extent of self citations as a measure of knowledge cumulativeness PUBKNOW: Extent of public research organizations and universities, over all HCITTECH:The concentration of technology classes among the cited patents

25. 4. Results Cluster analysis: 3 different clusters of technological domains Examples: Semiconductors/ Electronic commerce Examples: Traffic control systems/ Electricity storage & measurement Examples: SSD organic / Tech. for elec. & magn. measurement Little innovation by new entrants in a fast growing cluster => New knowledge acquired outside? => High level of opportunity Continuous opportunities for the entry of new actors Low degree of knowledge cumulativeness => Room for radical changes? 25

26. 4. Results Summary of Cluster Analysis Results • Green domains: • High dispersion of innovative activity across organizations, countries and technological classes. • Public knowledge plays a marginal role. • Degree of self-citations is low (low degree of knowledge cumulativeness & lack of specific firms’ competitive advantage). • Established domains : • Technological areas with a well-established innovative activity, which concentrates within few actors and countries. • Very high technological pervasiveness: their potential fields of applications spread across different technological areas. • Semiconductors and Electronic commerce are the most relevant domains in this cluster. • Emerging domains: (key findings) • Prevalence of ICT-related technological classes and companies + variety of knowledge sources, with an important role of universities and public research centres. • Innovative activity carried out by established firms through the acquisition of new competences outside their core area of technological expertise. • Represent areas with a high level of opportunity for future development. 26

27. 5. Implications for future research Conclusions about innovation dynamics in Green ICTs • Green ICTs cover 65 technological domains (different combinations of green & ICT classes). • Fastest growing domain = SSD using organic materials (Germany strongest EU country). • France: strong in Secure electronic commerce & Hall effect devices (thanks to CNRS). • LEDs = promising Green ICT: potential for job creation in Europe? (cf. FP7 cycLED project: http://www.cyc-led.eu). • Current set of technological classes in Green Inventory neglects important green ICT domains that stem from the combination of existing green and ICT classes. • 3 main clusters of green ICT domains (Green, Emerging, Established) that differ substantially in the structure of their innovative activity. • Further research: Green ICTs & relative/absolute decoupling? (magnitude of ecological benefits, job creation potential, ...). 27

28. Thank you. Grazia CECERE* Nicoletta CORROCHER# Cédric GOSSART* Müge ÖZMAN* # KITeS, Bocconi University, Milan * Télécom Ecole de Management, Institut Mines-Télécom, Paris ECOPATENTS project http://ecopatents.wp.mines-telecom.fr/

29. Extra slides A review of the ecological impacts of ICTs evaluated with life cycle analyses (in French, list of references –mostly in English- available from http://ecoinfo.cnrs.fr/article279.html): 29

30. Long term impacts of green ICTs: Absolute decoupling ICTs ? Well being Relative decoupling Absolute decoupling Natural resources time 30

31. Systemic changes towards absolute decoupling Source: Tukker, A. and M. Butter (2007). "Governance of sustainable transitions: about the 4(0) ways to change the world." Journal of Cleaner Production 15(1): 94-103. 31