PICMET 10 Conference PICMET ‘‘10 Conference Phuket, Phuket, Thailand Thailand 10R0142
Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective Lawrenzo Hung-Chun Huang Shin-Yu Shih Ya-Chi Wu National Chi Nan University, Taiwan
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Globalization! Global Supply chain, Global outsourcing , Global R&D
National Innovative Capacity
International Technology Diffusion
Global Network Structure
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
National innovative capacity National innovative capacity has been defined as the institutional potential of a country to sustain innovation and a suitable measure based on patenting rates.
Endogenous determinants Common Innovation Infrastructure
Cluster-Specific Environment
National Innovative Capacity
patenting rates
Quality of Linkages
Furman, J. L., M. E. Porter, et al. (2002). "The determinants of national innovative capacity." Research Policy 31(6): P906.
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
1.Endogenous fully determine ? Endogenous determinants
Exogenous affect
?
National Innovative Capacity
Can an endogenous perspective fully determine national performance in terms of innovative capacity?
Furman, J. L., M. E. Porter, et al. (2002). "The determinants of national innovative capacity." Research Policy 31(6): P906.
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
2.Exogenous environments? International Technology Diffusion
Global Network Structure
affect
National Innovative Capacity
Endogenous determinant s Furman, J. L., M. E. Porter, et al. (2002). "The determinants of national innovative capacity." Research Policy 31(6): P906.
What kinds of international relationships have a greater effect?
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
3.Exogenous influence; International Technology Diffusion International Technology Diffusion
affect
National Innovative Capacity
Endogenous determinant s Furman, J. L., M. E. Porter, et al. (2002). "The determinants of national innovative capacity." Research Policy 31(6): P906.
What difference of the alternate channels of ITD ? Are their differential impact on NIC?
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Related theory for new perspective 1.Network Autocorrelation Model (Leenders,2002) 2.The flows of international Technology Diffusion (Griliches, 1989) 3. Interactive effect, social contagion theory of innovation (Burt, 1986 etc.)
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
1.Network Autocorrelation Model (Leenders,2002) Simultaneously considers both endogenous determiners and exogenous influence on national innovative capacity. Global Networking Environments
Interactive Effects
Domestic Factor Individual diligence
Local Effects
National Performance Innovative Capacity
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
2.Flows of International Technology Diffusion Griliches,1980 Embodied Technology→ Rent Spillover ) eg. direct investment, trade flows (eg. Grossman & Helpman, 1991; Smith &White, 1992; Coe & Helpman,1995; Xu & Wang, 1999,2000; Eaton & Kortum, 2001. Nemeth &Smith, 1985; Kick & Davis, 2001)
Disembodied Technology→ Pure Knowledge Spillover )eg. licensing, outsourcing agreement, patent citations (eg. Griliches,1980; Austin,1993; Kong & Lin, 2003)
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
3.Interactive Effects, a major engine of diffusion network Cohesion mechanism (e.g.. Burt, 1991 Berelson et al., 1954; Koka et al. 1999, Shih, 2006) communication
Structural equivalence mechanism (eg. Galaskiewicz & Burt, 1991; Harkola & Greve, 1995; Koka et al., 1999; Shih, 2006)
comparison ego ego
alter alter
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Conceptual framework & Data collection Diffusion
Interactive Effect
Exogenous ITD
Intermediates
Technological diffusion
Communication
Cohesion
Comparison
Structural Equivalence
Embodies
Trade Flow
Disembodies
Patent Citation
Common Innovation Infrastructure Endogenous
Local Effect
Furman Model
Cluster-Specific Environment Quality of Linkages
National Innovative Capacity
Patents Output Data collection
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Data Collection ¬Samples: 42 countries
(form Global Competitiveness Index of the World Competitiveness Databank top 42 Countries)
¬Periods:1997-2005 ¬Variable: ) ) ) )
Innovative Output Æ International patents granted (form USPTO) Aggregate R&D Expenditure ÆTotal R&D expenditures (from IMD) Embodied technology Æ trade flows (from GTI) Disembodied technology Æ patent citations (from NBER)
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Result & Finding Discussion Part 2
Discussion Part 3
Cohesion
Structural Equivalence
Cohesion Embodies Tech.
Disembodies Tech.
NIC Performance Domestic
Discussion Part 1, 4
Structural Equivalence
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Result
Existing exogenous effect?
Endogenous determinants
NIC Performance
Furman’s Model Significant Support Endogenous determinants
R2=0.96
However…Exist serious Deletes variables with collinearity
muliticollinearity problem
Furman’s Model
insignificant Support Endogenous determinants
R2=0.50
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Result
Existing exogenous effect?
Adjusted R2
Embodied y Technolog
0.8 0.6 0.4
ied Disembod y Technolog
Furman Model
Co+SE
Cohesion
Deleted collinearity
Cohesion
Structural Equivalence
Co+SE
Fu ll M od el
1
Structural Equivalence
0.2 9 M od el
8 M od el
7 M od el
6 M od el
5 M od el
4 M od el
3 M od el
2 M od el
M od el
1
0
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Finding 1 Exogenous effects NIC ? Replace Furman Model’s collinearity variables with interactive variables. Global Networking
Interactive Effects Domestic Factor
Local Effects
NIC Performance
Autocorrelation Model
Significant Support
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Finding 2 What kinds of international relationships have a greater effect? Global Networking
VS.
Structural Equivalence
alter alter
Structural Equivalence
Cohesion
Cohesion
ego ego
+
Interactive Effects Local Effects
Cohesion Embodies Tech.
NIC Performance
Structural Equivalence
NIC Performance
+
Disembodies Tech.
NIC Performance
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Finding 2 What kinds of international relationships have a greater effect? Adjusted R2
ied Disembod y Technolog
Embodied y Technolog
0.8 0.6
Cohesion
0.4
Structural Equivalence
Co+SE
Cohesion
Co+SE
Fu ll M od el
1
Structural Equivalence
0.2 9 M od el
8 M od el
7 M od el
6 M od el
5 M od el
4 M od el
3 M od el
2 M od el
M od el
1
0
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Finding 3
What difference of the alternate channels of ITD ? Are their differential impact on NIC? Comparison for dis/embodied technology diffusion Embodies Tech.
Domestic
Rent spillover
Productivity efficiency
NIC Performance
Pure knowledge spillover
Disembodies Tech.
Technical change
Difference in spillover rigidity Æ“rent spillover” are more rigidity than “Pure knowledge spillover”
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Finding 3 What difference of the alternate channels of ITD ? Are their differential impact on NIC? Adjusted R2
i ed Disembod y Technolog
1
Embodied y Technolog
0.6
Co+SE Cohesion
Cohesion
Structural Equivalence
Fu ll M od el
0.8
Structural Equivalence
Co+SE
0.4 0.2 9 M od el
8 M od el
7 M od el
6 M od el
5 M od el
4 M od el
3 M od el
2 M od el
M od el
1
0
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Conclusion 1.national innovative capacity simultaneously determiner by local effect (domestics factor) and influence by interactive effect (global environments). 2.the important differences between the alternate channels of ITD on NIC show that disembodied technology diffusion more effectively replenishes the indigenous technology environments than embodied technology does. Therefore, this channel differential provides policy alternatives in national science and technology development.
3.In a global network context, differential interactive proximity effects differentiate national innovative capacity. • The cohesion proximity negatively affects NIC, merely utilizing the technology of a cohesion partner will less contribute for innovations. • The structural equivalence proximity significantly influence NIC. That is, countries become more inclined to take competitors as a paradigm via international technology diffusion based on the environment in which they are developing.
4.National innovative capability is more significantly affected by foreign disembodied technology. Consequently, acquiring competitor countries’ disembodied technology is more effective to influence on innovative capability gain.
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Thank you Hung-Chun Huang Shin-Yu Shih Ya-Chi Wu National Chi Nan University, Taiwan
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Countries of International Technology diffusion Argentina
Australia
Austria
Belgium
Brazil
Canada
Chile
China
Colombia
Denmark
Finland
France
Germany
Greece
Hong Kong
Hungary
Iceland
India
Indonesia
Ireland
Italy
Japan
Malaysia
Mexico
Netherlands
New Zealand
Norway
Philippines
Poland
Portugal
Russia
Singapore
South Africa
South Korea
Spain
Sweden
Switzerland
Taiwan
Thailand
Turkey
United Kingdom
United States
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Methodology & Hypotheses Methodology ¬International technology diffusion
ITD ij = wij × RD i
¬Contagion effects ⎛ ⎞ y i = ρ ⎜⎜ ∑ wij y j ⎟⎟ + ε j≠i ⎝ j ⎠
( )
yi = ρ y + ε * i
j≠i
PICMET’10 Constructing National Innovative Capacity in Globalization: The Network Autocorrelation Perspective
Collinearity statistics Dependent variable=(PATENTS)j,t+3 Collinearity Statistics Toleran ce
VIF
GDP per capita
0.560
1.787
GDP
0.012
84.574
R&D personnel (FTE)
0.368
2.716
R&D $
0.012
82.324