Governance, Sustainability and Pathways to Food and Agricultural Futures
Life cycle of technological Cluster
1. CLUSTERS FOR LIFE OR LIFE CYCLES OF CLUSTERS
In search for the critical factors of clusters resilience
Raphaël Suire
CREM-CNRS
University of Rennes 1
Jérôme Vicente
LEREPS
University of Toulouse 1 Capitole
& Toulouse Business School
2. BASIC QUESTION / SUMMARY
• The question
– The driving forces of clusters life cycles and resilience in turbulent
economic environments (financial crisis, new growing societal
paradigms, short life cycle in high-tech sectors)
• Definition
– According to the ecological metaphor (Holling, 1973), resilient clusters
decrease their vulnerability through their endogenous adaptive ability
to resist to exogenous shocks and stresses, by reorganizing themselves
towards new technological fields and markets
• Previous related works
– From critical factors of clusters stability (Suire, Vicente, Regional
Studies 2007, Journal of Economic Geography 2009) to critical factors of
their resilience (this paper)
• Results
– Resilient clusters have to mix audience and networks effects along the
phases of the composite technological process, and display
core/periphery (degree distribution) and dissortative (degree
correlation) structural properties
3. THE RESILIENT SILICON VALLEY
1
HP
21
Silicon Valley 150 Apple
41
ranking (sales) Intel
61 AMD
Blue: computers 81 LSI
black: semiconductors Google
101
red: Internet services Yahoo
121
Green: solartech E-bay
141
SunPower
from MercuryNews 161
1980s
181
1990s Hypothesis: The overlapping of knowledge
and business domains in the Valley arise more
from knowledge compositeness and
transversality than from a simple and
2000s fortuitous co-existence
4. THEORETICAL BACKGROUND
Location decision externalities
• Audience and network effects
• Marshallian vs. Jacobian externalities
Stability of
Clusters life Co-location
locational
cycles and process and
norms and
resilience technological
technological
critical factors relatedness
standards
Composite technologies life cycle Technological Structural properties of knowledge
• Phases of the knowledge value overlapping and networks
chain structural • Core/periphery
• Standards diffusion and competition conditions for • Assortativity
resilience
• The related mechanisms of clusters resilience
– Location decision externalities (Arthur, 1990; Appold, 2005;
Romanelli, Khessina, 2005; Suire, Vicente, 2007-09)
– Composite technologies lifecycle (Klepper, 1996;
Malerba, Orsenigo, 1996; Audretsch et al, 2008)
– Topologies of knowledge networks (Coleman, 1988; Burt, 1992;
Jackson, Wolinsky, 1996; Barabasi, Albert, 1999; Newman, 2003)
5. 3 PROPOSITIONS
• Proposition 1: Clusters life cycles depend, ceteris paribus, on
the nature of location decisions externalities that govern the
co-location process.
• Proposition 2: The endogenous resilience capabilities of
clusters are, ceteris paribus, a function of the particular
overlap and feedbacks between the knowledge phases along
the composite product life cycle.
• Proposition 3: The resilient properties of clusters are, ceteris
paribus, a function of their core/periphery structure and their
degree of assortativity (degree distribution, degree
correlation)
6. PROPOSITION 1
• Location decisions of predecessors influence decision of
followers, but according to different motives or different constraints
(Suire, Vicente, 2009):
– Legitimacy, reputation and cascade effects (audience effects)
– Knowledge accessibility effects (network effects)
• Geographical charisma and places reputation lead to unrelated
knowledge variety (R&D productivity signal), as well as cognitive
homophily (sectoral productivity signal), whereas micro-motives
based on external knowledge accessibility can lead to related
variety.
Results in literature: network effects engender more stability for the
aggregate structure than audience ones
7. PROPOSITION 2
• From idea to the market, knowledge processes
are going through successive exploration and
exploitation phases. Between both, for
composite technological products, an
integration phase occurs
Ө2
• Demand characteristics evolve along the
lifecycle: from ongoing development and non-
zero default to well designed and price
Ө1
competitive products or services.
Ө0 • Innovations cross the chasm and diffuse from an
early market (Ө1), to a mass market (Ө2), when
Exploration Integration Exploitation
the integration phase leads to a better
satisfaction of consumers and an increase of
demand.
• Considering the battle of technological
standards, successful clusters are the ones that
succeed in crossing this chasm.
8. PROPOSITION 3
• Sociology : Coleman, Burt and Granovetter: information and signals diffuse through
social ties. Social capital, a resource that you get from social network, is a source of
individual performance : trade-off between cohesion with natural homophilic
behaviour and openness (weak ties), fresh air and non redundant signals
• Economics : (Jackson & Wolinsky, 1996), if investment in a network is time, money or
learning costly, then a rational individual should compare benefit and cost for each tie.
• Complex systems : (Albert, Barabasi, 2002; Newman, 2003): Networks are not random
structures and exhibit regular collective properties. Among them:
– Scale free networks: preferential attachment and growth : the rich gets richer with
power law distribution of degree
The hierarchy or the distribution of influences within the network and
cohesiveness
– The assortativity and openess
the network is assortative (positive degree correlation): the higher degrees
interact with high degrees: the lowest degrees interact with low degrees
the network is dissortative (negative degree correlation): the higher degrees are
tied with low degrees, the lowest degrees interact with high degrees
9. PROPOSITION 3
– A random network, flat distribution of degrees, neither
assortative nor dissortative: difficult to reach the critical
mass for technological standardization
– A core-periphery network with assortativity : easy to
reach standardization but what next ? Lock-in and high
sensitivity to external shocks
– A core-periphery network with dissortativity : easy to
reach standardization but what next ? Lock-out thanks
to lower sensitivity to external shocks resilient
network
11. THE DECLINING CLUSTER
Ө0 Ө1
Beta test Early It does not succeed in crossing
market market the chasm and stays locked into
External audience and the Ө1 market due to:
reputation effects prevail
and knowledge homophily
• Irrational exhuberance around a
place or a technology : locational
cascade and mimetic behaviours
Declining clusters • Strong cognitive homophily with
an excess of competition and a
lack of coordination
• Strong knowledge diversity which
prevents from coordination
External audience effects
prevail and knowledge
between heterogeneous
heterophily profiles, which remain in isolated
strategy
• Rather flat hierarchy
12. THE DOMINATED CLUSTER
Ө0 Ө1 Ө2
Beta test Early Chasm Mass
market market market
• Thanks to the network effects along the phases of technology
development
• Endogenous development with sectoral and complementary spin-offs
as well as external actor locations. Connection between these
peripheral members and the core of initial innovators: cohesive
structure as a necessity to cross the chasm for a well-oriented
exploitation
• Deficient inter-cluster actors and geographical gatekeepers in charge
of world-wide coordination: behind the battle of standard, the
dominated cluster losts the battle of cluster
• A hierarchy occurs with assortativity but not enough to exhibits the
existence of global pipelines between major local hubs
13. THE DOMINANT CLUSTER
Ө0 Ө1 Ө2
Beta test Early Chasm Mass
market market market
• Quiet similar to the previous one regarding the evolutionary path
(strong cohesiveness) along Ө1 and Ө2 market except that gatekeeping
strategies are observed. Strategic alliances in world-wide markets
permit a global coordination on a very cost-oriented market strategy
• The life cycle of the cluster follows the life cycle of the technology but it
can decline with the demand decrease as soon as an assortativity of the
network and a lack of knowledge variety prevent from investment in
new emerging technological fields and markets
• Strong hierachy within the cluster with a limited peripheral connected
variety
14. THE RESILIENT CLUSTER
Ө0 Ө1 Ө2
Beta test Early Chasm Mass
market market market
• Quiet similar to the previous one except that now the cluster exhibits
dissortative property such that explorative behaviour and disruptive
ideas are connected to the core of the network : knowledge heterophily
• Necessity to have geographical gatekeepers (global diffusion) as well as
in-betweenn actors, the linkers between periphery and core, such that
overlapping practices and relational continnum co-exist.
• This cluster exhibits resilience properties as its life cycle is disconnected
from the technology life cycle
• Cattani & Ferriani (2008) provide a convinced example of such a
resilience for the Hollywood industry
15. CONCLUSION (1)
• Clusters will be resilient if they combine network effects (the
structuring of the technological field) and external audience effects
(potential variety for markets overlapping)
• Such a process is going through an evolving core/periphery structure
– reinforcement of the standard in a cohesive structure
– entrepreneurial connections with the periphery
• For that, network cohesiveness, which is a regular result in the
studies of collective behaviors, has to go with a certain level of
dissortativity (as in physical networks), in order to engender more
modularity and dynamical flexibility to the structure of knowledge
flows
• One of the reasons of the long and enduring success of the Silicon
Valley is certainly due to the capacity of its social networks to play
with mobility, dissortativity, cross-border knowledge, and beliefs on
entrepreneurship
16. CONCLUSION (2)
• Policy implications
– Most of clusters policies, and European policies in particular, are based on the
increase of the density of local knowledge interactions: “connecting people”
one-best-way
– A strict application of this basic principle can be under efficient (reinforcement
of the core, negative lock-in and opportunistic behaviors regarding subsidies)
– Need for more surgical and targeted interventions centered on particular
missing links Cluster failures arise more from a failure in the distribution of
knowledge interactions rather than a strict weakness of their density per se
– Need for diagnostic-based policies for more modular and dissortative
networks, that connect in a more flexible way the core of well-established
organizations to the myriad of burgeoning companies and science-based
organizations, rather than the core in itself
– Focusing on particular missing links rather than a general watering of public
funds for coordination is suited to “repair” the lack of connectivity some
clusters exhibit.