Réunion plénière du RT8 – 12/09/2013
Patterns of innovation in Green ICTs:
A patent-based analysis
Nicoletta CORROCHER#
Grazia CECERE*
Cédric GOSSART*
Müge ÖZMAN*
# KITeS,
Bocconi University, Milan
* Télécom Ecole de Management, Institut Mines-Télécom, Paris
An output of the ECOPATENTS project funded by ADEME
http://ecopatents.wp.mines-telecom.fr/
Outline
1. Introduction
2. Background literature
3. Methodology & Data
4. Results
5. Implications for future research
2
1. Introduction
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
3
1. Introduction
Source: Hilty (2008: 147).
ICTs & the environment
Green ICTs
Ecological impacts of ICTs
4
1. Introduction
Green ICTs
(hardware + software)
Reduce negative ecological impacts + create jobs:
1. Direct impacts: ICTs reduce their own ecological
impacts.
2. Enabling impacts: ICTs are used by other sectors to
reduce ecological impacts.
3. 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
2. Literature
Ecoinnovation & patents
1. Increasing number of studies using patents to analyse
ecoinnovation (cf. lit. rev. in our paper: http://ssrn.com/abstract=2117831).
2. The propensity to patent varies across sectors including for
eco-technologies (=> which ones are the most dynamic?).
3. Eco-innovation studies cover different technological
domains, e.g. automobile, waste, water, air cleaning (not ICTs).
6
3. Methodology & Data
Objective & Questions
 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
7
3. Methodology & Data
Data
Construction of our green ICT technological domain:
1.
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).
2.
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
3. Methodology & Data
Methodology
3 steps:
1. Select EPO patents that have at least one green and one ICT
technological class (at 7 digit level) granted between 1987
and 2006.
2. 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.
3. We analyse network components (technological domains) in
detail.
3. Methodology & Data
Methodology
1. Select EPO patents that have at least one green and one ICT
technological class (at 7 digit level) granted between 1987 and 2006
3. Methodology & Data
International Patent Classification (IPC) codes
Green
ICT
IPC codes
IPC codes
a, c
3. Methodology & Data
International Patent Classification (IPC) codes
Green
IPC codes
b
Pure Green
ICT
IPC codes
ICT
IPC codes
3. Methodology & Data
Methodology
80
H01S 5/34
H01L 33/00
ICT Code
H01S 5/323
H01S 5/34
H01S 5/343
H04L 9/32
H04L 12/58
H04L 29/06
H04L 29/06
69
Green Code
H01L 33/00
H01L 33/00
H01L 33/00
G06Q 20/00
G06Q 10/00
G06Q 10/00
G06Q 30/00
109
80
85
62
69
67
80
67
80
2. Network analysis:
Analyse the structural characteristics of this network.
Clusters of codes can define technology domains.
4. Results
Network analysis – Detection of Green ICT domains
65 different technological fields (components):
4. Results
The components of green ICT technological fields
15
4. Results
Patents & IPC classes in green ICT fields
Technological component
Number of
Patents
Number of
IPC classes
ICT classes
Green
classes
Pure Green ICT
Other classes
Semiconductors
2058
1248
33%
2.4%
4.7%
59.9 %
Arrangements for testing electric
properties and locating electric faults
2849
1085
34.8%
2.3%
4.7%
58.2%
Secure electronic commerce
669
406
48.1%
1.7%
0.2%
50%
Arrangements or instruments for
measuring magnetic variables
705
370
7.3%
2.4%
4.6%
85.7%
Nuclear magnetic resonance
technologies
658
160
12.5%
2.5%
10%
75%
Solid state devices using organic
materials
470
489
9.6%
1%
0
89.3%
Traffic control systems
899
335
38.8%
1.2%
2.4%
57.6%
Hall effect devices
284
126
55.5%
0.8%
5.5%
38.2%
Liquid crystal displays
209
643
1.7%
0.2%
0
98.1%
Technologies for total factory control
75
77
23.4%
3.9%
0
72.7%
Excitation or detection systems
360
68
20.6%
0
13.2%
66.2%
16
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
4000
0
2000
cleanpatentnb
6000
most of the patents).
0
20
component
40
60
17
Annual growth rate of patents in Green ICTs (1987-2006)
4. Results
Average annual
growth rate =
0,34 (34%)
18
4. Results
Firms with the highest number of Green ICT patents
Firm
Number of patents
SIEMENS BUILDING TECHNOLOGY AS
IBM
HEARTSTREAM
HEWLETT-PACKARD
BUDERUS HEIZTECHNIK
KONINKLIJKE PHILIPS ELECTRONICS
ADVANCED CERAMICS
CANDESCENT TECHNOLOGIES
NIPPON TELEGRAPH AND TELEPHONE
FUJIKIN
FUJI ELECTRIC
TEKTRONIX
AT & T
MATSUSHITA ELECTRIC INDUSTRIAL
FUJITSU
CENTRAL JAPAN RAILWAY
AGILENT TECHNOLOGIES
ADVANCED MICRO DEVICES
ERICSSON
857
694
341
339
323
272
271
250
237
211
208
191
190
189
188
169
163
143
139
CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRS
123
19
4. Results
Top 3 organisations by Green ICT
technological component
Technological component
Total n° of
organisations
Semiconductors
537
Arrangements for testing
electric properties and
locating electric faults
785
Secure electronic commerce
315
Arrangements or instruments
for measuring magnetic
variables
191
Nuclear magnetic resonance
technologies
166
Solid state devices using
organic materials
141
Traffic Control Systems
253
Hall effect devices
101
Liquid crystal displays
39
Number of
patents
CANDESCENT TECHNOLOGIES
120
NIPPON TELEGRAPH AND TELEPHONE (NTT)
112
Top 3 organisations
IBM
IBM
SIEMENS BUILDING TECHNOLOGY AS
HEWLETT-PACKARD
IBM
FUJITSU
FRANCE TELECOM
ADVANCED CERAMICS
HEARTSTREAM
SIEMENS BUILDING TECHNOLOGY AS
ADVANCED CERAMICS
HEARTSTREAM
KONINKLIJKE PHILIPS ELECTRONICS
EASTMAN KODAK
CAMBRIDGE DISPLAY TECHNOLOGY
3M INNOVATIVE PROPERTIES
BUDERUS HEIZTECHNIK
SIEMENS BUILDING TECHNOLOGY AS
AISIN AW
IBM
BUDERUS HEIZTECHNIK
CNRS
CELANESE
CHISSO
F. HOFFMANN-LA ROCHE
105
122
160
128
36
28
22
103
38
36
68
64
42
52
35
33
109
71
51
44
14
12
63
36
14
20
4. Results
Top 3 innovative countries in Green
ICT technological domains
Technological domain
Semiconductors
Arrangements for testing
electric properties and
locating electric faults
Secure electronic commerce
Arrangements or
instruments for measuring
magnetic variables
Nuclear magnetic resonance
technologies
Solid state devices using
organic materials
Number of patents
Country
716
704
177
1151
472
329
329
220
130
345
89
66
329
81
62
170
109
74
US
JAPAN
GERMANY
US
JAPAN
GERMANY
US
JAPAN
FRANCE
US
UNITED KINGDOM
JAPAN
US
UNITED KINGDOM
SWITZERLAND
US
JAPAN
GERMANY
21
4. Results
Top 3 innovative countries in Green
ICT technological domains (cont.)
Technological domain
Traffic control systems
Hall effect devices
Liquid crystal displays
Technologies for total factory
control
Excitation or detection
systems, e.g. using
radiofrequency signals
Number of patents
Country
295
JAPAN
252
GERMANY
115
US
92
US
50
JAPAN
44
GERMANY
89
36
GERMANY
HUNGARY
24
UNITED KINGDOM
25
JAPAN
17
GERMANY
13
US
173
US
53
UNITED KINGDOM
48
SWITZERLAND
EU
22
4. Results
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).
23
Growth: The average annual growth rate of the domain wrt the number of patents
4. Results
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
Variable
Obs
Mean
Min
Max
252.2
818.4
.5082421
.4917579
.3475701
.0350057
.0281049
Std.
Dev.
435.0559
1850.121
.0411233
.0411233
.1669447
.0378638
.0299088
Number of patents
Number of technological classes
Share of ICT classes
Share of green classes
GROWTHi
ENTRYi
HORGi
65
65
65
65
65
65
65
6
110
.4182306
.2599558
.0971514
.0099174
.0055423
2849
11687
.7400442
.5817695
.8044047
.2595041
.2123942
HCOUNTRYi
65
.2016453
.0537544
.1264003
.4248914
HTECHi
65
.0192141
.0151048
.0041719
.0966121
SELFKNOWi
PUBKNOWi
HCITTECH
65
65
65
.0720965
.0038859
.0396431
.0339153
.0095034
.0277021
.0216138
0
.0014406
.1959799
.0527638
.1556497
24
Cluster analysis: 3 different clusters of
technological domains
4. Results
Variables
CLUSTER 1
(25 domains)
GREEN
CLUSTER 2
(11 domains)
EMERGING
CLUSTER 3
(29 domains)
ESTABLISHED
Average nb of patents
Average nb of tech. classes
Share of ICT classes
Share of green classes
172.32
236.60
.503
.497
128.09
207
.522
.478
368.14
731.04
.508
Examples: Semiconductors/
.493
Electronic commerce
.649
.212
GROWTHi*
Examples: Traffic control systems/
.372
Electricity storage & measurement
ENTRYi*
.030
HORGi*
.022
HCOUNTRYi*
.190
Pervaniseness_TECHi
.040
=> New
knowledge acquired
.053
.034 outside?
SELF_CITATIONii*
.062
=>.074
High level of opportunity
.080
ACADEMIC_KNOWi*
Variery_CITTECH*
.023organic / Tech. .044
Examples: SSD
for elec. & magn. measurement
.020 Continuous
.036
Little innovation
by new opportunities
entrants in a for
fast growing
the entry
cluster
of new actors
.206
Low degree.001
of knowledge
cumulativeness
.017
.210
.009
.004
.013
.024
=> Room for radical changes?
*Indicates statistically significant differences across clusters.
25
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
Conclusions about innovation
dynamics in Green ICTs
5. Implications for future research
1. Green ICTs cover 65 technological domains (different combinations of green &
ICT classes).
2. Fastest growing domain = SSD using organic materials (Germany strongest EU
country).
3. France: strong in Secure electronic commerce & Hall effect devices (thanks to
CNRS).
4. LEDs = promising Green ICT: potential for job creation in Europe? (cf. FP7
cycLED project: http://www.cyc-led.eu).
5. Current set of technological classes in Green Inventory neglects important green
ICT domains that stem from the combination of existing green and ICT classes.
6. 3 main clusters of green ICT domains (Green, Emerging, Established) that differ
substantially in the structure of their innovative activity.
7. Further research: Green ICTs & relative/absolute decoupling? (magnitude of
ecological benefits, job creation potential, ...).
27
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/
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
Long term impacts of green ICTs:
Absolute decoupling
Relative decoupling
ICTs
?
Absolute
decoupling
time
30
4(0) ways to change the world." Journal of Cleaner Production 15(1): 94-103.
Source: Tukker, A. and M. Butter (2007). "Governance of sustainable transitions: about the
Systemic changes towards absolute decoupling
31