Cluster Dynamics

in Theory and Practice

with Application to Scotland

Ross Brown

Regional and Industrial Policy Research Paper

Number 38

March 2000

Published by:

European Policies Research Centre

University of Strathclyde
40 George St.
Glasgow G1 1QE
United Kingdom

ISBN: 1-871130-16-6
Abstract

The paper reviews the academic and policy literature on clusters and cluster-based
development strategies. In particular, it highlights the key issues which will affect
cluster development activities in Scotland and how Scottish Enterprise can progress
with its current and future cluster action plans. The paper has the following
objectives: to examine the theoretical issues surrounding clusters and cluster-based
strategies, including different definitions of clusters; to review the spatial levels at
which policy-makers attempt to implement cluster strategies; to assess the different
approaches and methods of analysing and measuring clusters; to examine the
evolution of clusters and how different clusters relate to one another; to highlight
issues involving ‘cluster governance’ and how different countries/regions structure
cluster programmes; and to provide some brief conclusions on how SE can take
forward its clusters approach.

Dr Ross Brown
European Policies Research Centre
University of Strathclyde

The author can be contacted at:

Scottish Enterprise
120 Bothwell Street
Glasgow G2 7JP
Scotland

2
TABLE OF CONTENTS

1. The Theory and Practice of Clusters................................................................. 4

2. Geographical Issues Surrounding Clusters ...................................................... 6

3. Measuring and Selecting Clusters ..................................................................... 8

4. The Evolution of Clusters ................................................................................. 10

5. Cluster Governance........................................................................................... 12

6. General Policy Conclusions .............................................................................. 13

Annex 1: Definitions ………………………………………………………………..15

Annex 2: Porter's Diamond ………………………………………………………..16

Annex 3: Cluster Policies by Country …………………………………………….17

Annex 4: Cluster Analysis Techniques ………………………………………...…26

Annex 5: Glossary …………………………………………………………………27

Annex 6: A Typology of Scotland's Pilot Clusters ………………………………28

References …………………………………………………………………………29

3
1. The Theory and Practice of Clusters
This section examines some of the theoretical issues surrounding clusters and cluster-
based policies. It is fair to say that there has been a large degree of confusion
surrounding the concept of clusters and that clusters or clustering means different
things to different people. One of the reasons for this is the fact that there is no real
unified theoretical framework for examining clusters. Feser (1998) notes there is no
cluster theory per se, rather a broad range of theories and ideas that constitute the
logic of clusters. Some argue that this situation has had negative policy ramifications:
‘sadly, in the rush by various governments to employ clusters, some fundamental
issues have been slighted, including appropriate research methods and even the
definition of the cluster itself’ (Held, 1996, p.249). Held claims that this approach can
be harmful.

Part of this problem owes to the fact that there has been a variety of different
definitions used when examining clusters1. Michael Porter defines clusters as
‘concentrations of interconnected companies and institutions in a particular field’
(1998, p. 78). Meanwhile, the DTI (1998, p.22) defines clusters as ‘a concentration of
competing, collaborating and interdependent companies and institutions which are
connected by a system of market and non-market links’. Scottish Enterprise (SE)
define clusters as ‘customers, suppliers, competitors and other supporting institutions
such as universities, colleges, research bodies, financial institutions and the utilities’
(Scottish Enterprise, 1998).

In practice, clusters are taken to mean a group of business enterprises and non-
business organisations for whom membership within the group is an important
element of each member firm’s individual competitiveness. Binding the cluster
together are ‘buyer-supplier relationships, or common technologies, common buyers
or distribution channels, or common labour pools’ (Enright 1997, p. 191). It is also
important to note that clusters involve a certain degree of spatial proximity between
its actors. Geographical proximity enables face-to-face networking, common labour
markets and the diffusion of knowledge, especially ‘tacit’ knowledge which is
difficult to codify.

Although these definitions have a large amount of commonality, not everyone views
clusters the same way however. For example, in the UK there is now a widespread
acceptance in policy-making circles that clusters are a ‘good thing’. From a national
perspective, clusters are seen as an effective way of promoting national economic
competitiveness across the UK as a whole through a process of regional or local
specialisation (DTI, 1998). From a regional or local perspective, clusters are viewed
more generically as a process for stimulating regional/local economic development.
Indeed, for local economic development practitioners clusters have rapidly moved
from being an academic term used to describe industrial agglomerations to a process
of economic development to promote local business competitiveness and enhance
growth within local economies. For the purposes of this paper we will concentrate on
clusters in the latter sense.

1
Adding confusion to this is the fact that related terms, such as industrial districts, value chains and
business networks, are sometimes used interchangeably with the term clusters (see Annex 1).

4
Much of the upsurge of interest in clustering was sparked by the influential work of
Michael Porter (1990). It is widely recognised that most cluster studies use Porter’s
work as a starting point for cluster analysis (Bergman and Feser, 1999). Porter’s
model of national competitiveness - the so-called ‘diamond’- is by now well know
and warrants only a brief summary here (see Annex 2 for more details). Although
Porter’s diamond provides policy-makers with a useful starting point for examining
clusters, not all cluster initiatives proceed with this approach when deciding which
clusters to focus policy support upon. In fact, according to a study examining cluster
initiatives across Europe, the majority of cluster programmes were not preceded by an
in-depth regional analysis comparable with Porter’s cluster mapping process
(Lagendijk, 1999a). Moreover, Lagendijk claims that the general Porterian concept of
clusters has been interpreted very differently across regions and translated into
practical initiatives according to the specific needs of the local economy. This has
been backed up by other cluster researchers2. A study by Michael Enright (2000) also
found that very few clusters nowadays adopt Porter’s approach.

Most observers acknowledge, however, that Porter’s work is merely the starting point
for the majority of cluster studies whilst tapping into a wider range of more developed
ideas to explain the origins of industry clusters, the dynamics of cluster growth and
change and advantages to using clusters as a basis for regional policy (Bergman and
Feser, 1999). Current thinking on clusters lead by Michael Enright suggests that
Porter’s work should be used as a starting point or a catalyst to undertake cluster
development and not as some kind of manual which has to be rigidly followed. From
a policy perspective, different clusters are often highly place-specific and public
policy towards clusters should be flexible enough to accommodate varying industrial,
institutional and political conditions.

According to Feser (1998), there are broadly two different types of cluster policies.
First, policy applications that focus specifically on identified clusters called ‘cluster-
specific strategies’. Under the cluster-specific policy approach, the objective is to
encourage the emergence or development of a distinct identified cluster. The first
stage of this policy approach is to map out the cluster and identify its characteristics
and nature through SWOT analysis. The primary characteristic of the cluster-specific
approach is the comprehensive attempt to nurture a given value-chain through a range
of carefully crafted demand-side and supply-side policy interventions. Cluster-
specific approaches are mutually reinforcing and may involve economic interventions
which are not development strategies per se, but are instead traditional functions of
governments (eg. regulation, enforcement, pricing and education).

The second approach is termed ‘cluster-informed strategies’. The principal policy

objective from the cluster-informed perspective is the improved implementation of
individual development initiatives. This approach to cluster development is less
holistic than the cluster-specific approach. No formal cluster mapping exercise is
undertaken under this approach, but some cluster-related techniques may be deployed.
Cluster-informed strategies often seek to examine specific aspects of a cluster. For
example, such an approach might seek to examine parts of a clusters supply chain (eg.
weaknesses in local suppliers) which would then be targeted by business development

2
The initial meeting of Euro-Cluster also found this to be the case.

5
initiatives. Under this approach clusters are primarily viewed as an analytical device
to improve the effectiveness of narrower types of policy tools (Feser, 1998).

Annex 3 outlines different approaches taken towards clusters across countries

throughout Europe. It is clear from this analysis most countries fall somewhere
between these two camps outlined above and significant inter-national (eg. Sweden
and Finland) and inter-regional variations (eg. Wales and Scotland) arise in the
implementation of cluster-based development strategies across Europe. With the
exception of the Netherlands and parts of Scandinavia, we can see that few countries
have an explicit overarching national cluster policy (Boekholt and Thuriaux, 1999).
Nonetheless, many countries and regions adopt policies which are clearly ‘informed’
by notions such as clustering and networking (see Annex 3). What is also interesting
is that some countries not examined within, such as the Republic of Ireland, who have
examined a clusters approach to economic development then opt for a more
traditional sectoral approach (see NESC, 1998). These differences clearly reflect the
fact that different countries have vastly different political structures and systems for
national industrial policy, science and technology and regional development.

2. Geographical Issues Surrounding Clusters

This section examines the issues which affect the operation of cluster policies across
different spatial and institutional scales. There are a number of general issues
concerning the geography and scale of clusters that policy-makers should take into
consideration when assessing how cluster policies might need special tailoring to
address issues facing their regional economies (OECD, 1997). These include:

I. Geography: A cluster approach should consider such factors as geographic

size and proximity to other important regional economies.

II. Jurisdiction: Jurisdictional issues affect assessments of regional economies

as many regional economies are composed of multiple jurisdictions (ie.
municipal; provincial/state; federal) which possess their own sub-regional
economies and public policy regimes3.

III. Population: Population size can be significant as it can determine the size of
local labour forces and local consumer markets. Additionally the urban/rural
split can be a further policy factor in regional economies.

IV. Sector Size: Many industrial sectors (and therefore potential industrial
clusters) are not particularly large in many regional economies. As a result,
any methodology used to measure or select clusters should be adaptable
enough to capture the key aspects and group dynamics of smaller industrial
groupings of like companies.

In addition to the factors mentioned above, the most appropriate scale for working
with clusters is also determined by other factors. For example, a recent DTI (1999)

3
For example, Porter (1998) notes that a strong chemicals cluster in Germany crosses the border into
neighbouring Switzerland.

6
enquiry into the biotechnology sector discovered that perceptions of cluster proximity
vary between countries. In the US, clusters tend to be thought of as locations that can
be visited within a single business day, and from this perspective the UK as a whole
might be viewed as a single cluster. In contrast, the DTI study team felt that in the
UK a cluster is more likely to be circumscribed by a much shorter journey (around an
hour or so). The geographical size of any given cluster is also determined by labour
market issues. In particular, how far staff are prepared to move their place of work
without moving house. Given this, clusters are often geographically circumscribed by
the perceptions of participants within the cluster, both businesses and workers.

Differences also arise in the width and breadth of clusters which are selected by
policy-makers. For example, in Northern Ireland the clusters chosen are very wide
ranging. Examples include broad industrial/service clusters such as ‘engineering’ and
‘tradable business services’. According to observers, such wide-ranging clusters are
too general to have much coherency for policy support (Enright, 2000). Meanwhile,
in Scotland a more narrowly focused approach has been taken with its first wave of
clusters. For example, SE has chosen to breakdown the electronics industry into
smaller sub or micro-clusters comprising elements of different sectors, eg.
semiconductors and opto-electronics rather than information industries/electronics as
a whole.

Although clusters tend to be geographically concentrated (see Porter, 1998),

determining the most appropriate scale to work with clusters is very important, not
least because clusters often cut across international, national, regional and local
boundaries. As we can see in Annex 3, there are significant variations in the
geographical scale at which countries/regions implement cluster policies. While some
countries operate a highly systematic top-down approach to cluster development (eg.
the Netherlands), in some other countries cluster policies are implemented by local or
regional actors in a very ad hoc way (Scandinavia and to a lesser extent the UK).
Determining the most appropriate spatial scale for adopting cluster policies is
obviously a key issue for policy-makers and will normally be influenced by the
political and institutional structure of any given country.

In countries with a large degree of political or economic autonomy, such as Spain and
Germany, cluster policies are typically undertaken by local or regional actors. In fact,
cluster policies are sometimes most developed in regions which have the largest
degree of autonomy (eg. The Basque Country and Catalonia in Spain). The reason for
this owes to the well developed economic development infrastructure which is found
in highly decentralised countries. For example, in the UK it is probably no
coincidence that the regions which have been quickest to embark upon cluster-led
economic development are those with the highest levels of political and institutional
autonomy. In particular, Scotland has been at the forefront of clusters during the
1990s through the work of Scottish Enterprise.

Cluster development strategies are sometimes devised at the national level but are
then implemented locally by a different set of actors. This then creates problems
when implementing local cluster policies. How can central agencies ensure that local
cluster policies are properly implemented? When this situation occurs it is important
that close coordination takes place and appropriate coordination mechanisms are in

7
place between different economic agencies at various spatial scales4. Very little, if
any, research has been undertaken to assess how different countries/regions surmount
these difficulties. However, it would appear that very close linkages and information
flows between different sets of actors will be needed to ensure this works effectively.

3. Measuring and Selecting Clusters

Although numerous methods are used to analyse and measure the size and importance
of regional clusters, there is no universally accepted method of cluster assessment and
measurement (see Held, 1996). Different countries and regions tend to define clusters
in a variety of ad hoc ways, using a wide variety of research techniques and criteria.
According to some observers, ad hoc procedures are always questionable for defining
and implementing policy (Doeringer and Terkla, 1995). Indeed, there is a large
cluster ‘tool box’ from which clusters can be identified: for example, industry-based
input-output relationships, shift-share analysis, location quotients, industry growth
forecasts, case studies and predictions about the business potential of particular
products and technologies (Rosenfeld, 1995). The variety of different cluster
analytical techniques also owes much to the different techniques which are deployed
by various economic development bodies and consultancies working in this growing
field. Cluster identification methodologies also differ in the use of economic analysis
and techniques (see Annex 4 for a summary).

One of the main problems which can hinder cluster analysis is the lack of available
regional data, a problem which is particularly acute in Europe. Porter (1998) claims
that clusters rarely conform to standard industrial classification systems. This
problem means that standard economic analysis may fail to capture many important
actors and relationships. Even in the US, which has a relatively disaggregated
statistical nomenclature (ie. 8-digit SIC), problems identifying clusters can occur. For
example, in Massachusetts more than 400 companies representing at least 39,000
employees, are involved in medical devices in some way yet statistical analysis failed
to detect this sizeable cluster. This can sometimes lead to a situation whereby policy-
makers and businesses do not recognise their existence owing to overlapping industry
categories such as electronic equipment and plastic products. Despite this,
representatives from this ‘hidden’ cluster have now come together to tackle common
problems. Some countries, such as Denmark, are currently undergoing a process of
re-classifying their statistical classification systems so that clusters are better captured
by official data sources.

In addition to these measurement problems, Sternberg (1992) warns against an overly

mechanistic approach towards cluster identification because such an approach may
fail to recognise rapidly emerging or embryonic clusters which have not yet reached a
critical mass and may require support from public policy to maximise its potential5.

4
As we shall see in section 5, this problem is sometimes complicated when business actors are brought
into the process of ‘cluster governance’.
5
A good example of this would be the optoelectronics cluster in Scotland which would not be ‘picked
up’ by conventional cluster measurement techniques. This mini-cluster has substantial growth
potential but needs proactive policy support if its potential is to be achieved. Similarly, had Finland
relied on retrospective economic data during its cluster analysis during the early 1990s, the

8
Unlike SE’s approach, few countries pay little explicit attention to industry growth
forecasts when selecting clusters to focus policy efforts on. However, one of the
reasons why the so many regions select so-called ‘fashionable’ clusters probably owes
to the projected growth rates of these industries.

Practitioners generally favour the use of both quantitative and qualitative analysis
towards cluster identification and analysis (Held, 1996). This is the approach which
the DTI is taking to its current research which is attempting to map industrial clusters
across the UK. Following consultation with external experts, the DTI were advised to
adopt a ‘nested’ or ‘mixed’ methodology when measuring clusters. This approach
will make use of various official data sources such as the Inter-Departmental Business
register (IDBR) as well as close liaison and interviews with local business leaders,
RDAs etc. Such an approach attempts to marry ‘hard’ quantitative analysis with
‘softer’ forms of qualitative understanding6. Clearly some clusters are more difficult
to define than others.

Following the process of cluster analysis and identification, policy-makers are then
faced with the difficult decision of deciding which clusters to work with. The process
of selecting clusters is one of the most controversial aspects facing economic
development practitioners when adopting a cluster-based approach and care must be
taken during the selection process so that clusters which do not get given priority by
development agencies are not then neglected (Enright, 2000). Lagendijk (1999b)
claims that there are two principal routes to cluster selection:

1. The first is the ‘top-down’ approach in which a list of clusters is drawn up on the
basis of cluster analysis. This approach faces the problem of justifying why
certain clusters were included and others were excluded. Enright (2000) claims
that policy-makers often find it difficult to resist ‘fashionable’ clusters (eg.
biotechnology, multimedia etc.) even when there is little genuine background in
these areas. Even if this temptation is resisted, the process of cluster selection
will always be based on certain arbitrary choices. This approach is found in
places such as the Basque Country (Lagendijk, 1999b).

2. The second is the ‘bottom-up’ approach which sees the initiative being taken by
actors within the sector themselves. Bottom-up approaches generally benefit from
existing inclinations to undertake joint action and to cooperate. This type of
approach involves a process of self-selection and may only suit strong clusters
with existing patterns of horizontal/vertical cooperation, ignoring weak clusters
with strong development potential. This approach is common in southern
European countries such as Spain and Italy, especially the Italian industrial
districts.

telecommunications cluster would not have been included in its list of clusters for policy support (see,
Rouvinen and Ylä-Anttila, 1999).
6
The latter approach is particularly important when deciding which clusters policy-makers should
focus upon.

9
In reality, the majority of clusters selected for policy help and support involve a
combination of these two routes7. For example, in the state of Arizona local
governments offer clusters the opportunity to bid for public funds to undertake
clustering initiatives, ensuring that only the clusters with existing cooperation were
likely to win funding. Meanwhile, in Scotland a combination of the two approaches
also took place owing to the fact that extensive private sector consultation took place
prior to launching a clustering approach in each of the four pilot clusters.

4. The Evolution of Clusters

As many observers have noted, clusters are not a static phenomenon. Once a cluster
begins to form, a self-reinforcing cycle promotes growth, especially when local
institutions are supportive and local competition is vigorous (Porter, 1998). Vibrant
clusters are constantly innovating, changing shape and altering their internal dynamics
(Baptista, 1998). In fact, the reconfiguration of any given industry or cluster is often a
mark of the level of innovation within its component parts. Indeed, clusters which
remain quite stable and do not transform themselves may end up stagnating,
especially if they do not upgrade and keep abreast of new technology. There are
numerous examples of regions which have been severely hit because a dominant
industrial cluster failed to diversify, upgrade or move up the value-chain (shipbuilding
in Glasgow being a good example).

Although different clusters have varying trajectories, some seem to develop along a
life-cycle model (Swann, 1998). Often the life-cycle of any given cluster owes to
new developments in technology and when a new technology or innovation process
occurs, new or embryonic clusters emerge. These often arise from research within
universities and research bodies. A good example of this in the Scottish context is the
emergence of an opto-electronics cluster which is now beginning to grow and
develop, primarily from the Scottish research base. Given this cluster is made-up of
university researchers and small spin-off companies, this cluster is clearly in the
formative stages of its development and will probably not fully develop until well into
the next century. In fact, according to Porter (1998), numerous case studies suggest
that clusters require a decade or more to develop depth and real competitive
advantage.

Clusters which follow the above type of developmental trajectory are often formed by
commercialising the existing science base. Using an example for a hypothetical
cluster, the life-cycle could be categorised as follows:

7
Lagendijk (1999b) claims that a mixture of the two approaches is probably the best method of cluster
selection.

10
1. Stage 1: The research base creates the technology which forms the bedrock for
the future development of the cluster.

2. Stage 2: The research base commercialises the technology either internally

through spin-offs or via a third party contractors.

3. Stage 3: Companies begin to grow and develop around the technology creating
a small cluster of related firms. Employment levels are still quite low.

4. Stage 4: Specialist support services and suppliers emerge to service the cluster,
adding to the depth and innovative capacity of the cluster as a whole.

5. Stage 5: Companies begin to grow and expand their capabilities. Often trade
bodies or informal networks grow during this stage.

6. Stage 6: Companies begin to reach a degree of maturity and may need to

diversify to be able to continue in business. Linkages with universities or other
suppliers of technology may be necessary to upgrade.

Not all clusters of industrial activity follow this simple linear developmental path.
Often clusters arise as a consequence of convergence between traditional and newer
technologies. One of the most widely quoted examples is the convergence of
telecommunications and information technology. Some computing and electronics
companies start to specialise in the branches of technology required to make this
interaction happen (modems, communication cards, digital exchanges etc.).
Following this, companies in one industry start to diversify into another. For
example, the telecommunications giant Nokia has used its wide-ranging experience in
telecoms to move into other forms of ICT which were traditionally seen as outwith its
core business. Indeed, the convergence between traditional technologies and new
ones can sometimes yield substantial dividends (Swann, 1998).

Just as clusters may have a varying life-cycles or developmental trajectories, there is

also the need for public policy towards clusters to adapt over time to the needs of a
cluster at any given time. Therefore, policies appropriate to one stage of the cluster
life-cycle may not be appropriate to another. Thus, from the hypothetical example
quoted above, policy-makers such as Scottish Enterprise will have to ensure that
appropriate policies are implemented at any given stage of a cluster’s development.
For example, the early stages of cluster development require help with the
commercialisation process. Often the early phase requires pro-active hand-holding on
behalf of public agencies. This can take the form of assistance with market research,
testing, prototyping and general marketing. Following this companies within the
cluster may require assistance gaining venture or patient capital to aid their expansion.
During the later stages, economic development assistance may be confined to
attracting specialised foreign investors, improving networking in the cluster and
developing ‘bridging mechanisms’ between industry and the research base.

However, it is important to stress that the needs of clusters will vary and there is no
uniform ‘tool-kit’ which can foster cluster development. Furthermore, appropriate
policies will also be required for ‘existing’ or ‘mature’ clusters. For example, much
of the policy focus for existing clusters, particularly mature clusters such as food and

11
drink, is geared towards fostering networking and developing better technology-
transfer mechanisms. Finally, the need for a customised approach is even more vital
for clusters which are built upon foreign investment. These transplant clusters often
require a completely different set of policy instruments to aid their development, such
as pro-active supplier development, incentives to facilitate new R&D facilities and
help to develop the managerial autonomy within branch plants in order to gain
secondary investments.

5. Cluster Governance
This section highlights some of the issues surrounding ‘cluster governance’.
According to Michael Enright, a cluster’s governance structure refers to ‘the
relationships among firms in the cluster in terms of the nature if relationship and
distribution of power’ (2000). Cluster governance typically refers to the structure of
the industry and how firms interact with one another. Coordinating mechanisms, on
the other hand, refer to the ways that inter-firm relations are organised and how the
public sector seeks to influence the development of the cluster. We shall now
examine in these two features in turn.

A clusters governance structure is comprised of a variety of different aspects (see

Annex 5). For example, the activity base of a cluster involves the number and nature
of the activities in the value-added chain that are performed within the region. In
activity-rich clusters, most or at least many of the critical activities in the value-added
chains of the relevant industries are performed locally. Activity-poor clusters, on the
other hand, involve one or only a few activities in a given industry or set of related
industries. Within Scotland, examples of activity-poor clusters would be certain parts
of the electronics industry, especially those that are predominantly foreign-owned,
which perform little value-added locally. Other elements of cluster governance, such
as cluster density and breadth, are outlined in Annex 5.

Applying Enright’s framework to Scotland’s pilot clusters is a good way of

illustrating some of the dynamics of cluster governance. As we can see in Annex 6,
Scotland’s three pilot clusters are quite different in their structure and overall
dynamics. Therefore, SE will have to undertake policy interventions which are suited
to each of the cluster’s characteristics and needs. Some of the key characteristics are:

1. The biotechnology cluster is composed of a small number (sparse) of firms with a

diverse range of activities (broad) with no leading or core firms which are spread
across Scotland (dispersed).
2. The semiconductor cluster is characterised by a small number (sparse) of lead
firms with a narrow range of activities (narrow), mostly based in Central Scotland
(localised).
3. The food and drink cluster comprises of a large number of firms (dense) in a very
diverse range of businesses (broad) with no leading firms, spread across all of
Scotland (dispersed).

Coordinating mechanisms refer to the role played by different institutions in

developing clusters. For example, in some parts of the world clusters are coordinated
primarily by private sector. This is the case in many parts of the US and some parts

12
of Europe. In Arizona, for example, different clusters bid for public funding and are
then in charge of making interventions which will benefit the rest of the cluster. This
ensures that members within the cluster are pro-active enough to work together to
establish mutual goals and objectives. This type of approach ensures industry ‘buy-
in’ but may only be appropriate for clusters which are already well-developed and
have good internal linkages. However, in other parts of the world, such as Scotland,
economic development agencies may have to take the lead.

According to Enright, characterising clusters along these dimensions allows us to

understand their potential and problems in ways which can inform policy.
Knowledge of governance structures and coordinating mechanisms can also guide
policy towards the most efficient use of scare resources, especially as clusters, even in
the same location, might have very different characteristics. In general, policy-
makers should try and help localise, deepen, broaden, activity enrich and/or improve
the innovative capacity of clusters. From a policy perspective policies directed
towards SMEs will be somewhat ineffective in a cluster, such as semiconductors in
Scotland, if core firms are not interested in the programmes operated by SE.
Similarly, SE will have to act as the main coordinating mechanism, at least in the
short-term, for cluster policy if industry does not take up the challenge.

6. General Policy Conclusions

The clusters approach has emerged as an important analytical tool for governments
and economic development agencies seeking policy prescriptions to make their
economies and firms more competitive. While the cluster approach has traditionally
been used for examining national economies it can also be a useful tool in analysing
the dynamics of sub-national or regional economies. However, it is absolutely vital
that cluster programmes and actions are properly tailored to the individual needs and
requirements of any given cluster and the specific characteristics of any given region.
It is important, therefore, to stress that the needs of clusters will vary and there is no
uniform ‘tool-kit’ which can foster cluster development

This paper has highlighted a number of different issues which can help public policy
when adopting a cluster’s perspective. The most significant of these conclusions for
Scottish Enterprise were:

1. There is no overriding cluster theory per se and clusters are generally viewed as a
economic development process rather than a definite development theory.
Approaches towards developing clusters vary significantly across different
countries with some adopting a holistic cluster-specific approach while others
adopt a less far-reaching cluster-informed approach. SE should avoid adopting a
fixed attitude to ‘cluster thinking’ and should instead think laterally about the
specific needs of Scotland’s clusters. Working closely with external organisations
(eg. Scottish Executive, SHEFC, Coshep, SFEU, trade bodies) should help to
ensure that SE’s work on clusters also helps to influence other organisations and
policies so that cluster policy becomes fully embedded throughout Scotland.

13
2. Given clusters often transcend different jurisdictions, care must be taken to ensure
that close coordination takes place between different actors, especially when
cluster strategies are devised at a national level and then implemented at a local
level. This is something Scottish Enterprise will have to ensure now that the
implementation of the cluster action plans transfers from SE to other parts of the
Network. SE could also try and increase the role played by Highlands and Islands
Enterprise in the clusters process, especially in clusters which clearly transcend
both their jurisdictions (eg. tourism and forestry). This could be undertaken by
having representation from HIE’s at SE’s cluster events and/or inviting HIE’s
sector teams to become involved in SE’s cluster Implementation teams.

3. It is important that cluster methodology is carefully tailored to address the special

characteristics of the locale being examined and a mixed approach to cluster
measurement seems the most appropriate for bodies such as SE. Therefore,
Scottish Enterprise should try and be flexible when assessing which current and
future clusters to focus policy support upon. It is also important to recognise that
clusters change very rapidly, not least because the rapidly changing nature of
technology, and thus require a constant redefinition and focus. When it comes to
selecting which clusters to focus policy support upon, care should be taken to
ensure that members of the cluster are actively involved from the beginning in a
‘bottom-up’ manner. This reinforces the need for extensive discussion and
development work with members of clusters prior to devising cluster strategies. It
is therefore important that SE do not assume a fixed attitude when defining
clusters which may omit potential or emerging clusters.

4. Given clusters have varying life-cycles or developmental trajectories there is a

need for public policy towards clusters to adapt over time to the needs of a cluster
at any given time. Owing to rate of change within existing clusters, especially
high-technology business clusters, policy-makers will have to be able to track and
examine cluster dynamics in order to be aware of future technological
developments which can positively and negatively affect clusters. One method of
undertaking this is to engage in ‘foresight exercises’ which can help steer the
future of existing clusters towards new growth opportunities by improving
industry awareness and channelling research resources appropriately (Swann,
1998).

5. Recognising different cluster governance structures and coordinating mechanisms

can help to guide policy towards the most efficient use of scare resources,
especially as clusters, even in the same location, might have very different
characteristics. By taking the lead on cluster policies, SE is then in a position to
establish public-private sector bodies which can then take up the lead for future
cluster-related activities.

6. Finally, SE should try and help localise, deepen, broaden, activity enrich and/or
improve the innovative capacity of clusters.

14
Annex 1: Definitions

Sector (or Industry) A sector or industry is a group of enterprises that

manufacture similar products, as typically defined, for
example, under the U.S. Standard Industrial
Classification (SIC) system.
Regional industry cluster A cluster whose elements share a common regional
location, where region is defined as a metropolitan
area, labour market, or other functional economic unit.
Potential industry cluster A group of related and supporting businesses and
institutions, that, given additional core elements, inter-
firm relationships, or critical linking sectors, would
obtain some pre-defined critical mass.
Value-chain industry A value chain cluster is an industry cluster identified as
cluster an extended input-output or buyer-supplier chain. It
includes final market producers, and first, second and
third tier suppliers that directly and indirectly engage in
trade. It is comprised of multiple sectors or industries.
(see Roelandt and den Hertog 1999).
Business network ‘A group of firms with restricted membership and
specific, and often contractual, business objectives
likely to result in mutual financial gains. The members
of a network choose each other, for a variety of
reasons; they agree explicitly to cooperate in some way
and to depend on each other to some extent. Networks
develop more readily within clusters, particularly
where multiple business transactions have created
familiarity and built trust (Rosenfeld 1995a, p. 13).’
Ties between firms in networks are typically more
formal than in clusters.
Italian industrial district A highly geographically concentrated group of
companies that ‘either work directly or indirectly for
the same end market, share values and knowledge so
important that they define a cultural environment, and
are specifically linked to one another in a complex mix
of competition and cooperation between firms, a result
of a close intertwining of economic, social, and
community relations.
Innovative milieu Not a group of business or a region, but a complex
which is capable of initiating a synergetic process, an
organisation, a complex system made up of economic
and technological interdependencies, a coherent whole
in which a territorial production system, a technical
culture, and protagonists are linked.

Source: Based on Bergman and Feser (1999)

15
Annex 2: Porter’s Diamond

Firm Strategy,
Structure and
Rivalry

Factor Demand
Conditions Conditions

Related and
Supporting
Industries

Porter traces the competitiveness of firms to four major national factors:

1. the nature of firm strategy, structure, and rivalry in the country, including attitudes
toward competition, market institutions, the degree of local competition, and other
cultural and historical factors affecting how firms do business with each other,
their workers, and the government;
2. factor conditions, or basic endowments or conditions on which local firms seek to
compete (eg. such as supplies of natural resources pr technology-related advanced
factors);
3. demand conditions or the nature of local demand (eg. the existence of local
demand for related intermediate goods); and
4. the presence of related and supporting industries, including suppliers and
successful competitors; both to stimulate cooperation and competitive rivalry.

Source: Porter, 1990

16
Annex 3: Cluster policies by country
This annex examines the nature of cluster implementation in different parts of the
Western Europe. This survey is not meant to be an exhaustive list of cluster policies,
rather it is designed to act as an indicative overview of the types of policy approaches
towards cluster development in a selection of countries/regions in Western Europe.

Austria (cluster-informed)

Cluster-based economic development policies are quite common throughout Austria,

especially in certain regions with regional industrial specialisms. As with all
industrial and regional policies in Austria, cluster policies are part of a system of
‘social partnership’. This implies that basic policy principles and the conception of
programmes are agreed upon in advance in an informal cooperation between business
(Chambers of Commerce), labour (Trade Unions) and the state (federal and Land)
(see Todtling and Sedlacek, 1997). However, owing to the high degree of political
decentralisation, cluster policies vary markedly across Austria according to regional
and local needs. The main actors taking forward cluster development initiatives are
regional governments and regional development agencies.

At the national-level, one of the most ambitious plans for translating the cluster idea
into new policy is the ‘Competence Centre Programme’ (see Peneder, 1999). This
pilot programme was first initiated in 1997 by the federal government. The primary
goal of this programme is to improve interaction between academic research and
private industrial R&D through the establishment of collaborative research
institutions, called ‘competence centres’. Another cluster-informed policy initiative
in Austria is the creation of a new type of technical college specialising in a range of
activities from information and communication technologies such as engineering
skills and tourism management (Peneder, 1999).

One of the most proactive regions undertaking cluster-based policies is Styria. The
Styria automobile cluster is probably Austria’s best-know regional cluster. Building
on high performing local enterprises in motor vehicle technology and gear units, a
successful cluster of related companies has developed in this south-eastern province
of Austria. The cluster currently has more than 120 companies, as well as a host of
research institutes and technical colleges. Styria has used the cluster concept to create
a ‘brand name’ in order to attract foreign direct investment in the automotive supply
chain. The cluster concept has also helped to foster cooperation among local
suppliers, ranging from transport logistics to the new development of particular
automotive components (Peneder, 1999).

Belgium (cluster-informed)

In Belgium, national cluster policy is still in its infancy, although there are signs that
the concept is beginning to be given a higher priority by national and regional
government (Roelandt and den Hertog, 1999). Nonetheless, several policy initiatives
are strongly informed by cluster approach, such as the support for cross-sectoral
technology diffusion, the creation of supplier-producer networks and centres of

17
excellence around newly emerging technologies. At present, the main actors involved
in cluster development in Belgium are regional governments and RDAs. For
example, the RDA in the Kempen region is currently implementing the PLATO
programme. This aims to increase knowledge transfer between SMEs and large
international companies. The Plato programme brings together a group of SMEs from
different sectors in thematic sessions with a mentor company from a large firm. A
group of some 12 SMEs and two mentors participate in the groups for two years. The
emphasis of the programme is to share knowledge on management issues. This
initiative has been in operation since 1988 and has been copied by several other
European countries. Although not essentially a cluster initiative, this programme
teaches firms, especially SMEs, the benefits which can be obtained from working
with other firms (Boekholt and Thuriaux, 1999).

France (too early to know)

As mentioned previously France is just beginning to examine the potential of adopting

a national approach towards cluster development. The main national body in charge
of cluster work is DATAR and they have identified 230 different clusters located
across the country. This process has been undertaken in a very centralist fashion and
suggests that cluster selection procedures will be very ‘top-down’. At the regional
level, a number of RDAs are also beginning to examine the merits of adopting a
cluster-led approach to economic development. For example, Nord-Pais-du-Calias
and the Alsace region are both undertaking some basic exploratory work in this area.
However, it is too early to really assess how this will progress.

Germany (cluster informed)

Germany has a long tradition of collaborative research and networking, without

having an explicit cluster focus to its economic strategy (Boekholt and Thuriaux,
1999). Indeed, many of its R&D policies are designed to encourage research
collaboration. More recently, cluster-oriented policies have been developed by the
federal government. For example, in order to promote networking and clustering in
the biotechnology sector, the German government launched the ‘BioRegio’ initiative
in 19968. This initiative requires regions to submit ideas for the development of
biotechnology on a regional basis and awards financial and other special support to
the selected regions. The main criteria for the selection of a region is the existence of
collaboration between all the parties concerned (ie. industry, universities and the
public sector). The regions selected so far are Munich, Rheinland, Rhein-Neckar-
Traingle and Jena.

The BioRegio fund aims to promote the transfer of scientific knowledge in

biotechnology from universities to German industry, thus facilitating the
commercialisation of biotechnology research into products and processes. In
particular, the scheme supports:

8
The BioRegio programme is now being considered as a model for other clusters in Germany (eg.
transport technologies and logistics).

18
• Cooperation between research institutions and medical centres in the area of
research and testing;
• Help in management issues for start-up biotechnology firms;
• A network to exchange experiences within workshops and conferences; and
• Privileged funding from federal research funds.

As with Austria, Germany’s federal system of government means that each Lander
has a high degree of autonomy to pursue its economic policies. Various cluster
initiatives are, therefore, actively pursued by regions across Germany, most notably in
North Rhein Westfalen (NRW), Barvaria and Baden Wurtemburg. NRW, for
example, has been very proactive in promoting the development of its new media and
telecommunications clusters which is centred around Dusseldorf and Cologne. In
particular, industrial policy in NRW has established a so-called media initiative,
aiming at both fostering communication and supporting firms and consortia to realise
new media projects (Heinze et al, 1998).

The Netherlands (cluster-specific)

The clusters concept is seen as a central goal of economic policy in the Netherlands.
The main actor involved in promoting clusters at the national level is the Ministry of
Economic Affairs. Given the size of the country clusters are normally viewed on a
national level (Wever and Stam, 1999). In 1990, the Ministry commissioned a Dutch
‘Porter’ study to assess the economic strength of the economy as a whole and a
selection of clusters . Since then, mapping and cluster analysis studies have been
conducted and supported by the Ministry. In sectors such as telematics and
telecommunications, discussion platforms have been organised with industry to
discuss common bottlenecks and opportunities for the future. In the Netherlands,
clusters are also seen as key element of national technology policy and technological
foresight exercises have also been performed for various clusters. Indeed, most of the
measures put in place over the last five years either apply to industry in general or are
aimed at stimulating the development of technological innovation projects by
consortia of industry, research and end-users (Boekholt and Thuriaux, 1999).

In 1997, the Ministry of Economic Affairs published its first White Paper on cluster
policy: Opportunities Through Synergy: The Public Sector and Innovative Clustering
in the Market9. The White Paper distinguishes four main roles for the public sector
when promoting clusters:

• To create favourable framework conditions for industry and services in general;

• To act as a broker in identifying cluster opportunities by bringing together supply
and demand;
• To provide strategic intelligence for industry; and
• To operate as demanding customer in the provision of societal needs.

9
For more information check the Ministry for Economic Affairs website @
http://info.minez.nl/ezenglish/

19
The Netherlands also has a number of RDAs who are actively pursuing cluster-based
development policies, such as the LIOF (the Regional Development Agency for
Limburg). However, the Ministry of Economic Affairs is not fully aware as to the
extent and nature of these activities.

Scandinavia

Denmark (cluster-specific)

Denmark has been actively pursuing a clusters approach to industrial development for
a number of years using a cluster-specific approach. In particular, Denmark has
deployed the concept of ‘resource areas’ for a number of years. This approach seeks
to move away from traditional forms of industrial policy such as support for
individual businesses, to a system of fostering networks which will maintain and
increase the competitive position of entire industrial clusters. In principal the entire
economy was divided up into eight broad resource areas: services, agro/food,
construction, environment/energy, transport/communication, medical/health,
consumer goods, tourism/leisure and general suppliers.

As in the Netherlands, discussion platforms involving industry associations and the

Council for Industrial Development have been organised to discuss common
bottlenecks and opportunities for the future. Within each of the areas mentioned
above, the Danish Ministry of Business and Industry set up a dialogue with a group of
key people from industry, research, business associations and other related
organisation. Following this, cluster mapping exercises were performed for the entire
economy as a whole. A key to the success of the resource area approach is that it
requires the involvement of a number of Ministries in the implementation of dialogue
outcomes (Boekholt and Thuriaux, 1999). Five core areas for potential policy were
defined:

1. Government regulation

2. Access to knowledge

3. Access to capital

4. Interaction between the public and the private sector

5. Conditions for international competition

Sweden (cluster-informed)

There is little explicit acceptance of a clusters approach to economic development at

the national level in Sweden. Although Sweden undertook various cluster mapping
exercises, including the original Porter study, these have not, to date, had a significant
impact on its industrial or innovation policy (Boekholt and Thuriaux, 1999).
However, the Ministry of Industry, Employment and Communications and NUTEK
are currently examining the concept once again.

20
Where a clusters approach is being utilised in Sweden, it is within ‘bottom-up’
regional development strategies. For example, each county in Sweden has just drawn
up ‘Regional Growth Agreements’ (RGAs). These new growth agreements are a
clear sign of the Swedish government’s desire for greater decentralisation in the
regional policy. The growth agreements also reflect a growing desire on behalf of the
Swedish government that regional policy becomes more growth-oriented and less
concerned with alleviating regional disparities. Adopted in each region across
Sweden, the growth agreements aim to achieve improved cooperation and a
comprehensive approach between different policy areas and bodies which work with
measures promoting growth and employment within regions. The government wants
these agreements to foster greater cooperation between different policy areas so that
sector-specific central government expenditure maximises regional growth and
prosperity.

The growth agreements are drawn up by the local state bodies, in close collaboration
with other local partners such as representatives of the local business community,
chambers of commerce, municipalities, and county councils. Although the main
purpose of the agreements is to get local partners to agree on measures to be taken and
on the best way to coordinate state resources at the local level on a bottom-up basis,
the government also expects that national agencies (eg. NUTEK) will work together
to deliver these agreements. A recent evaluation of the growth agreements found that
many were broadly similar in nature and nearly all adopted a clusters approach when
undertaking a SWOT analysis of their local economies.

RDAs in Sweden are also implementing cluster-informed policy approaches. One of

the most pro-active actors in this respect is the East Sweden Development Agency.
The East Sweden region is a very dynamic region with a host of world-class research
facilities and a clusters approach is being applied to its inward investment and
commercialisation policies.

Finland (cluster-specific)

In Finland, cluster analysis and cluster-based policies have been applied at both the
national and regional levels and the clusters concept is now widely accepted in
Finland (Rouvinen and Yla-Antitila, 1999). The main driver promoting the concept
has been the national level, Ministry of Trade and Industry. A major national cluster
study during the early 1990s was conducted (Advantage Finland). This identified
strong clusters (forestry), fairly strong clusters (base metals and energy), potential
clusters (telecommunications, environment, well-being, transport and chemicals) and
defensive clusters (construction and foodstuffs). Many observers argue that
‘Advantage Finland’ provided a much needed new vision for the economy, reshaping
policy-making and practices nation-wide (Steinbock, 1998). In particular, this
national cluster study has encouraged regional and local industrial associations and
private businesses to take a more proactive role in improving their operating
environment (Rouvinen and Yla-Antitila, 1999).

Furthermore, recent government decisions have showed a willingness to increase

R&D funding to the sectoral ministries for various cluster programmes. There are

21
now cluster research programmes for telecommunications, foodstuffs, transportation,
environment, forestry and health care. Analysts conclude that the main impact this
transformation has had is on Science and Technology policy which has reoriented
towards a focus on developing science centres which focus upon Finland’s emerging
technological clusters (Rouvinen and Yla-Antitila, 1999).

UK

England (cluster-informed)

Previously, the UK showed little interest in cluster-based policies at the national level.
However, considerable interest is now being shown in the clusters approach. The
Labour government’s recent competitiveness White Paper makes explicit reference to
the potential role played by clusters in fostering both national and regional
competitiveness (DTI, 1998). Presently, the DTI has been the main player examining
the potential of clustering in the UK as a whole. Under the supervision of Lord
Sainsbury, Minister for Science, the DTI established a working group of policy-
makers and academics to examine the UK biotechnology sector and benchmark this
against the US biotechnology sector. The research identified ten critical factors that
encourage cluster development (see Table 1 below). It also made specific
recommendations for developing the UK’s biotechnology cluster (see DTI, 1999).

At the regional level in England, cluster initiatives are being implemented on a rather
piecemeal basis. Clearly, informal or cluster-informed initiatives are already evident
in parts of England (see Table 2). Most notably, Cambridgeshire has been very
proactive and dynamic in promoting a cluster approach to developing its local
industry, especially in the high-tech areas of software and biotechnology. Many of
these initiatives have involved the establishment of new institutions which bring
industry, venture capital, researchers and the public sector together. The government-
funded ‘Eastern Region Biotechnology Initiative’, for example, is viewed as being
highly successful at promoting networking within the local cluster (Keeble et al,
1999). So far, it has held a conference to promote the region and conducted a survey
of the cluster and its needs. This body received assistance under the DTI’s
Biotechnology Mentoring and Incubator Challenge.

22
Table 1: Factors Encouraging Cluster Development

Strong science base Leading research organisations: University departments,

hospitals/medical schools and charities
Critical mass of researchers, World leading scientist(s)
Entrepreneurial culture Commercial awareness and entrepreneurship in
Universities and research institutes,
Role models and recognition of entrepreneurs
Second generation entrepreneurs
Growing company base Thriving spin-out and start up companies
More mature ‘role model’ companies
Ability to attract key staff Critical mass of employment opportunities
Image/Reputation as biotechnology cluster
Attractive place to live
Premises and Incubators available close to research organisations
infrastructure Premises with wet labs and flexible leasing arrangements
Space to expand, Good transport links:
Motorways, Rail, International airport
Availability of finance Venture Capitalists, Business angels
Business support services Specialist business, legal, patent, recruitment, property
and large companies advisors, Large companies in related sectors (healthcare,
chemical, agrifood)
Skilled workforce Skilled workforce, Training courses at all levels
Effective networking Shared aspiration to be a cluster. Regional trade
associations. Shared equipment and infrastructure
Frequent collaborations
Supportive policy National and sectoral innovation support policies
environment Proportionate fiscal and regulatory framework
Support from RDAs and other economic development
agencies, Sympathetic planning authorities

23
Table 2: Sector/Cluster Orientation in UK Regions

Region Organisation Cluster focus Cluster methodology

Scotland Scottish Food and drink Focus groups, co-
Enterprise Oil and gas ordinating policy initiatives
Biotechnology along cluster lines (eg.
Semiconductors skills, technology
Wales Welsh Automotive sector, Supply chain initiatives;
Development consumer electronic links to centres of
Agency sector, medical excellence
devices and
diagnostics sector,
telecommunications
equipment sector
Northern Northern Engineering, food Supply chain initiatives;
Ireland Ireland processing, health creating new industry
Growth technologies, forums; developing better
Challenge software, textiles university-industry
and apparel, tourism linkages.
and leisure, tradable
services
North Northern Automotive, off- Varied top-down and
East Development shore, food, bottom-up initiatives, but
England Company: electronics, business primarily agency-led
Real Service services
Centre

Source: Adapted from Lagendijk and Charles (1999)

The onset of new RDAs in England will probably see an increase in regional cluster
development strategies. In fact, the DTI (1999) recommends that the new English
RDAs adopt a similar approach towards cluster development as in Scotland. In
particular, the working group advocated that the new RDAs look at improving the
environment for cluster growth, for example by addressing skills, planning, supply
chain and inward investment issues.

Wales (cluster-informed)

The Welsh Development Agency (WDA) has not officially embraced the cluster
approach to economic development, but some aspects of its sector-based policies are
very much akin to a clusters approach (see Table 2). In particular, the WDA has
concentrated its policy support on major sectors in the Welsh economy, most notably
automotive engineering and electronics and various emerging sectors such as
multimedia, medical, food and financial services (see Cooke, 1998; Lagendijk and
Charles, 1999). The main policy focus of the WDA has been to promote linkages
between foreign-owned firms and local suppliers in these sectors. Facilitating
technology-transfer between the HEI sector and business sector has also been a major
policy area. The WDA has played an important role at facilitating clustering in these

24
priority sectors by kick-starting the network process and building-up indigenous
supply chains, through forums, clubs and partnerships (Huggins, 1997).

Northern Ireland (cluster-informed)

The main body responsible for economic development in Northern Ireland is the
Department of Economic Development. During the mid-1990s, the Department of
Economic Development examined a clusters approach to economic development and
established Northern Ireland Growth Challenge (NIGC) in 1995. This body is a
private-sector led initiative operated by the Northern Ireland wing of the CBI in
conjunction with other social partners. It receives funding from the Department of
Economic Development and the EU under the special programme for Peace and
Reconciliation.

NIGC seeks to work with and develop the following clusters: construction, tourism,
food processing, engineering, software, health technologies, textiles and tradeable
business services. As noted previously, these very wide cluster areas have been
criticised by some (Enright, 2000). The main objectives of this body is to build
world-class firms/clusters, create networks of common interest, develop human
resources and market Northern Ireland to the world. Its main activities to date have
been fostering new networks of companies in these cluster areas. For example, within
the engineering cluster it helped to establish a Northern Ireland Aerospace
Consortium, which is a group of companies who cooperate to identify and develop
new business opportunities world-wide in the aerospace industry.

To date, nearly 200 companies have been associated with NIGC in different cluster
activities however, according to some observers within the Department of Economic
Development, NIGC has been somewhat disappointing regarding its progress with the
implementation of a genuine cluster-based approach in the province.

25
Annex 4: Cluster analysis techniques
The following are just some of the techniques which are utilised to measure and
assess clusters:

• Input-output analysis is used to identify trading relationships between different

actors within an economy. It is used to illustrate buyer-supplier relationships
within and between clusters. Regional economists use a range of methodologies
such as triangularization and factor/principal components analysis for sorting
industries into input-output linkages.

• Graph analysis, founded in graph theory, is similar to input-output analysis and

identifies cliques and other network linkages between firms or industry groups.

• The location quotient is the share of jobs that one industrial sector has in a region
in proportion to the sector’s share of all the jobs in the country or region as a
whole. Location quotients exceeding 1.25 are usually taken as initial evidence of
a regional specialisation in a given sector.

• Revealed comparative advantage measures the country’s share of exports from

each sector in relation to exports of all manufacturing sectors, and compared to the
average in the 13 OECD countries. National clusters are to be found in industries
with a high export rate, measured by the OECD’s revealed comparative advantage
index. For example, if a country has a revealed comparative advantage of more
than 8.5 in shipbuilding, the country has 8.5 times high a share of shipbuilding
exports as the average OECD country.

• The majority of cluster-related research into regional clusters/production systems

often deploys a case study approach or expert opinion approach. Regional
experts –industry leaders, public officials, and other key decision-makers- are
important sources of information about regional economic trends, characteristics,
strengths and weaknesses. Industry association reports, newspaper articles and
other published documents are also used when undertaking an expert opinion
approach. This approach is often utilised using Porter’s diamond approach
(Roelandt and den Hertog, 1999).

Each measurement technique, on their own, have limitations. Input output analysis is
only really useful for identifying clusters which are made up of buyer-supplier value
chains which include final market producers and first, second and third tier suppliers
that directly and indirectly engage in trade. The same criticisms also apply to graph
analysis. Location quotients say little about the inherent characteristics of regional
industry clusters. Furthermore, cluster studies that rely solely on location quotients to
identify clusters are simply sector studies in disguise because they offer no insight
into interdependencies between sectors (Feser, 1998). Revealed comparative
advantage is only useful in certain circumstances owing to the fact that data is often
only available for national rather than regional economies. Meanwhile, case study
analysis provides us with in-depth insight into the mechanisms which have created
various clusters and the methods of production used. However, case studies alone
cannot provide information about the quantitative importance of regional clusters.

26
Annex 5: Glossary
Organic cluster strategies seek to broaden and deepen a region’s existing economic
base.

Transplant clusters build clusters by attracting inward investment.

Dense clusters can be populated by hundreds of firms with total sales reaching
hundreds on millions/billions dollars.

Sparse clusters do not have the same economic weight, either because they have
fewer firms of fewer substantial firms.

Broad clusters provide a variety of products in closely related industries.

Narrow clusters consist of one of a few industries and their supply chains.

Deep clusters contain complete or nearly complete supply chains.

Latent clusters have a critical mass of firms in related industries, but have not
developed the level of interaction and information flows necessary to fully benefit
from co-location.

Potential clusters are those that have some of the elements necessary for the
development of successful clusters, but where these elements must be deepened in
order to benefit from the impact of agglomeration.

Policy-driven clusters are those which have developed primarily as a consequence of

government policy (eg. through the attraction of FDI).

Source: Enright (2000).

27
Annex 6: A Typology of Scotland’s Pilot Clusters

Cluster Biotechnology Semiconductors Food & Drink

Characteristics
Origin of Organic Transplant Organic
Industrial Base
Geographic Dispersed Localised Dispersed
Scope
Number/size Sparse Sparse Dense
and importance
of firms
Breadth & Broad Narrow Broad
Depth Shallow Shallow Deep
Innovative High Low innovation Low innovation
Capacity innovation
Cluster All ring, All core, All ring,
Governance No core No ring No core
(Research-led) (MNE-led) (SME-led)
Coordinating Public-Private SE (initially) SE (initially)
Mechanisms (BioAlliance)
Overall State of Potential Policy-driven Latent
Cluster
Development

28
References
Baptista, R. (1998) Clusters, Innovation and Growth, in Swann, P., Prevezer, M. and
Stout, D. (Eds) The Dynamics of Industrial Clustering: International Comparisons
in Computing and Biotechnology, Oxford University Press, Oxford.
Bergman, E. and Feser, E. (1999) Industrial and Regional Clusters: Concepts and
Comparative Applications, Regional Research Institute, West Virginia University.
Boekholt, P. and Thuriaux, B. (1999) Public Policies to Facilitate Clusters:
Background, Rationale and Policy Practices in International Perspective, in
Roelandt, T. and den Hertog, P (eds), Boosting Innovation: The Cluster Approach,
OECD, Paris.
Brown, R. (2000, forthcoming) Clusters and Local Embeddedness in Fyrstad,
European Urban and Regional Studies.
Cooke, P. (1998) Global Clustering and Regional Innovation: Systemic Integration in
Wales, in Braczyk, H., Cooke, P. and Heidenreich, M. (eds) Regional Innovation
Systems, UCL Press, London.
Doeringer, P. and Terkla, D. (1995) Business Strategy and Cross-Industry Clusters,
Economic Development Quarterly, Vol.9, pp.225-237.
DTI (1998) Building the Knowledge Driven Economy, Department of Trade and
Industry, HMSO, London.
DTI (1999) Biotechnology Clusters, Department of Trade and Industry, HMSO,
London.
Enright, M. (2000) The Globalization of Competition and the Localization of
Competitive Advantage: Policies toward Regional Clustering, in Hood, N. and
Young S. (eds) Globalization of Multinational Enterprise and Economic
Development, Macmillan, London.
Feser, E. (1998) Old and New Theories of Industry Clusters, in Steiner, M. (ed)
Clusters and Regional Specialisation, Pion Limited, London.
Heinze R., Hilbert, J., Nordhause-Janz, J and Rehfeld, D. (1998) Industrial Clusters
and the Governance of Change, in Braczyk, H., Cooke, P. and Heidenreich, M.
(eds) Regional Innovation Systems, UCL Press, London.
Held, J. (1996) Clusters as an Economic Development Tool: Beyond the Pitfalls,
Economic Development Quarterly, Vol.10, pp.249-261.
Huggins, R. (1997) Regional Competitive Specialization. Development Agency
Sector Initiatives in Wales, Area, Vol.29., pp241-252.
Keeble, D., Lawson, C., Moore, B., and Wilkinson, B. (1999) Collective Learning
Processes, Networking, and ‘Institutional Thickness’ in the Cambridge Region,
Regional Studies, Vol.4, pp.319-332.
Lagendijk, A. (1999a) ‘The Emergence of Knowledge-Oriented Forms of Regional
Policy in Europe’, Tijdschrift voor Economische en Sociale Geografie, Vol.90,
pp.110-116.
Lagendijk, A. (1999b) Good Practices in Cluster Development Initiatives: A
Summary of an ADAPT Funded Analysis of Cluster Initiatives, CURDS, University
of Newcastle Upon Tyne.
Lagendijk, A. and Charles, D. (1999) Clustering as a New Growth Strategy for
Regional Economies? A Discussion of New Forms of Regional Industrial Policy
in the UK, in Roelandt, T. and den Hertog, P (eds), Boosting Innovation: The
Cluster Approach, OECD, Paris.
NESC (1998) Sustaining Competitive Advantage, Proceedings of NESC Seminar,
Research Series, National Economic and Social Research Council, Dublin.

29
OECD (1997) Clusters and Regional Development Strategies: The Case of Atlantic
Canada, Working Party No. 6 on Regional Development Policies, Territorial
Development Service, OECD, Paris.
Peneder, M. (1999) Creating a Coherent Design for Cluster Analysis and Related
Policies: The Austrian ‘TIP’ Experience, in Roelandt, T. and den Hertog, P (eds),
Boosting Innovation: The Cluster Approach, OECD, Paris.
Porter, M. (1990) The Competitive Advantage of Nations, Free Press, New York.
Porter, M. (1998) Clusters and the New Economics of Competition, Harvard Business
Review, November-December, pp.77-90.
Roelandt, T., den Hertog P., van Sinderen, J. and Vollaard, B. (1997) Cluster Analysis
and Cluster Policy in the Netherlands, paper presented at OECD Workshop on
Cluster Analysis and Cluster Policies, Amsterdam, 9-10 October.
Roelandt, T. and den Hertog, P. (1999) Cluster Analysis and Cluster-Based Policy-
Making: The State of the Art, in Roelandt, T. and den Hertog, P (eds), Boosting
Innovation: The Cluster Approach, OECD, Paris.
Rosenfeld, (1995) Bringing Business Clusters into the Mainstream of Economic
Development, European Planning Studies, Vol.5, pp.3-23.
Rouvinen, P. and Ylä-Antitila, P. (1999) Finnish Cluster Studies and New Industrial
Policy Making, in in Roelandt, T. and den Hertog, P (eds), Boosting Innovation:
The Cluster Approach, OECD, Paris.
Scottish Enterprise (1998) The Clusters Approach: Powering Scotland’s Economy
into the 21st Century, Scottish Enterprise, Glasgow.
Steinbock, D. (1998) The Competitive Advantage of Finland – From Cartels to
Competition, Taloustieto Oy, Helsinki, Finland.
Swann, P. (1998) Clusters in the US Computing Industry, in Swann, P., Prevezer, M.
and Stout, D. (eds) The Dynamics of Industrial Clustering: International
Comparisons in Computing and Biotechnology, Oxford University Press, Oxford.
Todtling, F. and Sedlacek, S. (1997) Regional Economic Transformation and the
Innovation System in Styria, European Planning Studies, Vol.5, pp.43-63.
Wever, E. and Stam, E. (1999) Clusters of High Technology SMEs: The Dutch Case,
Regional Studies, Vol.33, pp.391-400.

30