1. Networks and innovation
Giuseppe Soda
Claudio Dematté Research Division
Prepared for Area Tec. Seminar, September 2009
Copyright SDA Bocconi 2005
2. Key Notion of Network
(relational) paradigm
The principal idea:
Economic action and
organizational performance are
influenced by the nature and the
structure of the relations between
and among organizational actors
3. Key Notion of Network (relational) paradigm
• Three perspectives:
– Network Analysis (social) as a Prism: networks as an analytical
device for illuminating relations, not necessarily social, whether
inside a firm, in the inter-organizational ties that link firms, or in the
environments of organizations;
– Ties and Structures of Ties as Explanatory concepts: network
theory (structural) argue that behaviors, actions, decisions, are best
discovered by examining relations among actors rather than just
their attributes;
– Networks as Org. Forms: it views networks as a kind of
organizing logic, a way of governing relations among economic
actors.
4. What is a network?
• A set of nodes (e.g. individuals, groups,
organizations) linked by a set of relationships
(Laumann et al., 1978)
• Relationships or ties can be
– Directed (one-directional)
– Undirected (as in being physically proximate)
– Dichotomous (present or absent)
– Valued (considering the strength)
5. Network perspective
in the management literature
• Network perspective builds on the general notion that economic
action does not take place in a solitary social context but it is
embedded within networks of interconnected relationships providing
opportunities and constraints on behaviour (Granovetter, 1973).
• Network perspective analyzes how the structure of relationships
influences performance and gives to the actors, such as individuals,
teams and organizations, the ability to acquire information, expertise
and other resources (Laumann et al., 1978; Fombrun, 1982;
Stinchcombe, 1990; Krackhardt, 1990; Burt, 1992; Ancona and
Caldwell, 1992; Brass and Burkhardt, 1993; Mizruchi, 1994; Powell
and Smith-Doerr, 1994; Ibarra, 1995; Podolny and Baron, 1997;
Brass, Galaskiewicz, Greve and Tsai, 2004; Rodan and Galunic,
2004).
• Network ties transmit information and are thought to be especially
influential information conduits because they provide salient and
trusted information that is likely to affect behaviour (Podolny, 2001).
7. Role of networks in innovation
• Through network ties and network structures organizations
can pool or exchange resources and develop new ideas
giving access to more diverse sources of information and
capabilities (1994; Powell, Koput and Smith-Doerr, 1996;
Obstfeld, 2005, Soda et al. 2004; Soda and Zaheer, 2009).
• Networks constitute a critical locus for knowledge creation
and innovation providing access to information, resources
and support necessary for new ideas (Powell and Brantley,
1992; Ibarra, 1993; Shan, Walker, and Kogut).
8. Networks and level of analysis
• Influence of network structures on innovation has been
investigated at
– Individual, team and organizational levels (Krackhardt, 1990;
Burt, 1992; Ancona and Caldwell, 1992; Brass and Burkhardt,
1993; Ibarra, 1995; Podolny and Baron, 1997; Rodan and
Galunic, 2004)
– Internal relations within teams and organizations (Gulati, 1998;
Rowley et al., 2000; Hansen, 1999, 2002; Tsai, 2001; Tsai and
Ghoshal, 2002; Gupta and Govindarajan, 2000; Levin and Cross,
2004)
– External relations among individuals, teams and organizations
(Ancona and Caldwell, 1992; Podolny, Stuart, Hannan, 1996).
9. Mixed empirical results on the relationship
between network structure and innovation
• Empirical evidence deriving from brokerage versus
closure argument on innovation is mixed
– Ahuja, (2000); Rowley et al., (2000); Obstfeld, (2005);
Uzzi and Spiro, (2005), Soda (2008) found at different
levels of analysis a positive impact of network closure
on innovation
– Baum et al., (2000); Ruef, (2002); Rodan and Galunic,
(2004); Zaheer and Bell, (2005); Fleming et al., (2007),
found a positive impact of brokerage on innovation
10. Exploiting structural position rents: Generating
innovation through strategic alliances networks
Giuseppe Soda
Copyright SDA Bocconi 2005
11. Strategic alliances
and firm performance
Alliance is a voluntary, interorganizational, formal
or informal agreement of cooperation or coordination
including exchange, sharing and joint development of
idiosyncratic assets, knowledge, and resources/
capabilities among the partners (Dyer, Singh 1998).
12. How does network position matter?
Usually managers:
• do not look at the alliance portfolio and whole network of
alliances in which they are embedded;
• they underestimate the impact of the positioning strategy
in the industry network of alliances
This research investigates the impact of network position on
firm performance.
More specifically we investigate to which extent product and
process innovation capacity can be affected by the position in
a network of strategic alliances
14. Research questions
• Does network centrality enhance
innovation outcomes?
• How network centrality makes influences
on innovation processes?
• Brokerage or closure?
15. Brokerage and Structural holes illustrated
Disconnections between neighbors and hence “entrepreneurial
opportunities for information access, timing, referrals, and control.”
Figure 1 - Illustration of Weak Ties and Structural Holes
1
Strong Ties
Weak Ties
1’s Ties to 2 and 3
Bridge Structural Hole
Structural Hole
2 3
16. The classic controversy between
cohesion and brokerage
• Coleman (1988): “closed • Granovetter (1973)
social structure engenders “ties within closed networks
trust because it enables are strong and connect you
collective sanctions” to similar others while weak
• Trust enables greater ties connect you to
sharing of resources, risk dissimilar others and fresh
and information info”
• Redundant ties facilitate • Cohesive networks recycle
information flow redundant information
• Easier coordination and • Brokerage provides power
mobilization and control advantages
20. Data
• 359 JVs 2001-2006
• 592 firms, 1136 links Figura 4: finalità delle Joint Venture
• Contents of Jvs: 60
– Assembling 50
– Commercial 40
30
– R&D 20
– Services 10
Production
Services
Commercial
– Production 0
2001
R&D
2002 Assembling
2003
2004
2005
2006
Assembling R&D Commercial Services Production
21. Data
• Patents (2001-2006) for all sampled firms.
Figura 9: numero brevetti e nuove alleanze
90 100
70 80
50 60
30 40
10 20
-10 0
2003 2004 2005 2006
Brevetti (migliaia) Nuove alleanze
22. Variables and analysis
Variables
Dependent variable:
Number of patents presented by the companies to the USPTO in the years 2004-2005-2006.
Although patents could not perfectly represent firms’ innovative capacity, they are a proxy used in several
studies (Powell, Koput, Smith-Doerr 1996; Hagedoorn 2002,; Ahuja 2000).
We considered presented and not registered patents in order to guarantee data homogeneity because the
time windows for the patents registration could be different for the several patents.
Independent variables
Cohesion is operationalized using the Density measure. Density is equal to (D = 2k / n*(n-1)), where k is
the number of existing ties among all the partners in the network and n is the number of partners in the
network.
Brokerage is operationalized using the Effective Size measure (Burt, 1992) that is equal to the number of
partners to which focal firm is connected minus the mean of partners belonging to the ego-network .
Control variables
Country is a dummy variable for the geographic area (Asia Pacific, Europe e North America)
Industry - Sic Code for the Production of complete vehicles (3711) and Production of components.
Analysis
We used Generalized Linear Model (GLM) selecting the Negative Binomial option to cope with the
overdispersion of the dependent variable (variance greater than mean)
23. Results
Patents 04
_____________________
b Std. Err.
Ego Density 01-03 -.01 .00
Structural Holes 01-03 .66 *** .15
Country Asia Pacific .95 *** .19
Country .88 ** .28
Sic 3711 -.04 .19
_cons 3.60 .20
Patents 05
_____________________
b Std. Err.
Ego Density 02-04 -.01 * .00
Structural Holes 02-04 .69 *** .12
Country Asia Pacific .92 *** .24
Country 1.55 *** .34
Sic 3711 .72 ** .22
_cons 3.22 .27
Patents 06
_____________________
b Std. Err.
Ego Density 03-05 -.02 *** .00
Structural Holes 03-05 .67 ** .22
Country Asia Pacific -.34 .29
Country Europe .34 .28 * p < .05
Sic 3714 .35 .20 ** p < .01
_cons 4.29 .31 *** p < .001
24. Conclusions
• Network position matters
Figure 1 - Illustration of Weak Ties and Structural Holes
• Bridging structural holes
enhance innovation
1
Strong Ties
• Network closure is Weak Ties
detrimental for
innovation 1’s Ties to 2 and 3
Bridge Structural Hole
Structural Hole
2 3