Clusters Are Obsolete

Are Clusters or cluster initiatives really adding any tangible value to firms? Are they obsolete? Maybe not yet, but I believe if the present policy tools and institutional frameworks are not reformed they may very well soon become so. In the following blog I would like to focus on three areas that have not been given enough importance by policy makers: focus beyond geographies, new clusters and cluster performance.

‘Clusters are geographic concentrations of interconnected companies, specialised suppliers, service providers, firms in related industries, and associated organisations (such as universities, standard agencies, trade associations) in a particular field linked by commonalities and complementarities. There is competition as well as cooperation.’ This is how Michael Porter defined clusters in the 80′s making them fashionable.

Others talked about them long before Prof. Porter. For example Marshall in 1920′s defined clusters as a concentration of a large number of similar businesses in a locality. The geographical aspect hasn’t really changed until now. Guiliani (2005) defines clusters as a ’geographical agglomeration of firms operating in the same industry. A lot has been said about how this agglomeration impacts economic growth. Nonetheless subscribing to Martin and Sunley’s critique of clusters (2003); I would argue there exists no direct causality.

Today however in practice, focus is more on innovation and for the same reasons on knowledge intensive industries in the policy making and academic circles. Clusters should, it is argued, help strengthen the Research and Innovation System of countries and regions (European Commission’s project Innova for example).  The fact remains however, that in the age where globalisation and technology has virtually made spatial concentration irrelevant, policy makers are still clinging to the cluster tune in its old form. I would argue that policy makers have to set different priorities in order to benefit their regions or countries more; enabling firms that actually drive innovation, appropriate value from these programs in terms of real radical or disruptive ideas. There are three main focus areas which are presently not given ample importance in the present discussion and action on clusters in the policy circles:

  1. Forget geographies- think bigger!
  2. Tracking new clusters is as important as encouraging old ones!
  3. What about cluster performance?

In the following, I would like to elaborate on the above.

1. Forget geographies- think bigger!

In a cluster, there exist subclusters prior to any institutional allignment based on the relationships intrinsic to the value chain. A subcluster that is a network of suppliers, customers, research organisations, media agencies and sometimes even competitors. The interconnections are based on the strategic intent, value expectations determined by complementarities, knowledge, power, size and position on the value chain; as well as trust. Trust in fact lubricates any network, reducing transaction costs. Countries that lack governance and institutional frameworks; family conglomerates with interests in various industries try to compensate for the same. Trust takes time to develop. Especially in terms of innovation that lies to the core of many firms; companies are not usually ready to open their doors easily. Hence, cluster institutional frameworks should act (and hopefully they also act) to connect these ’trust agglomerates’ or subclusters inside a cluster in a region.

Networking, matchmaking sessions or research seminars should at least in principle increase the probability for firms outside one’s subcluster to connect. We know however from social network theory that networks where there exist no structural holes (where if A knows B and also all nodes B knows are networks of strong ties), enhance the qualitative aspect of interconnectedness and frequency of contact in these networks and you increase the so called density of ties. In such networks, cognitive distance fades away and potential for exploratory or radical innovation is substantially lower as compared to that of incremental innovation (which is higher). That implies in fact that in any cluster the potential for radical innovation is limited and should further reduce with time as well as activities performed by cluster managers. Research also confirms that substantial radical ideas are produced on the boundaries of the industries rather than inside industries themselves.

This means focus should be rather laid on connecting what I call spatially and facultively distant clusters (firms in those clusters!). Policy should enable institutional frameworks enabling information sharing on not only geographically distant clusters and their firms of the same industry (Cluster of Clusters- a spatial Mega cluster); but also in faculty distance (Cluster of Clusters- an inter industry Mega cluster). We need to look at the cutting edge industries that lie on the fuzzy cross roads of different industries and build large ‘knowledge clusters’  transgressing different geographies and industries. That’s where in practice most of the radical innovation is happening anyway. This requires considerable political and institutional capital as local interests always tend to converge in building ‘local innovation engines’. It is archaic for companies today not to embrace collaboration as Brahmanical ivory towers of innovation are already history.

2. Tracking new clusters is as important as encouraging old ones!

In defining clusters, we talk of economic agglomeration. Usually for designing policy tools to enable economic growth; the importance of employment data is overplayed. What about capital intensive industries or creative industries? An animation studio might employ a handful of people but may require a billion dollars of investment. Spin offs, divestments by large or small corporations or investment flows usually set the scene to establish a value network.

The pace of innovation is significant as the base is low and the network is not ‘set’.  It is usually the start of an era of a creative destruction and it is also the time when the firms in such a nascent network require most of the assistance- distribution partners, knowledge partners, development partners- partners of all kinds. However, they don’t usually get assistance as the cluster institutional frameworks fail to track or recognise such clusters at a very early stage.

We need to establish policy tools and evolve present frameworks in order to enable the same.

3. What about cluster performance?

Do we really measure performance of a cluster in a way that acts as a feedback to the policy tools increasing innovation productivity? How many radical innovations could be attributed to the cluster framework? How many incremental innovations could be attributed to the same? What constitutes performance of a cluster manager? How do the activities performed by a cluster manager link directly to the policy goals (hopefully they do!) of increased innovation output?

We talk of regions in Europe whose innovation output is high. Well, innovation output was high in these regions probably even before they were labelled as ‘clusters’. In practice, companies search partners to collaborate -eventually hoping to innovate. Capability information arbitrage is necessary but does not need an elaborate institutional mechanism (and investment) in this age of digital connectivity.

It is argued that the investors invest in regions that drive innovation and hence one needs to assess the innovation productivity of a region. I believe investors primarily invest in opportunities or companies with specific technological or market capabilities. Even if the argument is true, we don’t observe any comparable and credible innovation capability assessment or innovation performance assessment performed at a regional/sectoral level that might be of interest to potential investors.

We need an inside out perspective on expectations of individual firms (potential firms and firms that already exist in the cluster)  to define the innovation agenda set by clusters but a perspective that’s democratic in nature. If we stake political capital in getting perspectives from firms, we might end up asking only the large and powerful ones. Essentially those who have failed radical innovation almost for a decade and not those who are the torchbearers of Schumpeter- the SME’s of today!

Are clusters obsolete? – probably no but in their present form they will be!

About the author

Gunjan Bhardwaj is senior editor and a member of the review team at Innovation Management. Gunjan is presently with the Boston Consulting Group and just prior to this he was the leader of the Global Business Performance Think-tank of Ernst&Young. Gunjan is also a guest professor for Growth and Innovation management at European Business School (EBS) in Germany and a member of the scientific advisory board of Plexus Institute in the US which researches on complexity in health sciences. Gunjan has published a number of papers and articles in various Journals and magazines and has been a frequent speaker in conferences on marketing and innovation related topics. The views and ideas expressed by Gunjan on are strictly personal and have no bearing on BCG.

  • Piero Formica

    My view is similar to yours. Take the case of clusters in Italy, dubbed industrial districts, which I have highlighted in my book INDUSTRY AND KNOWLEDGE CLUSTERS,
    Tartu University Press, 2004
    Piero Formica

  • haydn

    Porter was very explicit in saying you’d expect geography to have very little influence these days (and this was 1990, before the web). But surprise, surprise! It is still important. Then the surprise was that business was free from factors like coal and iron, even oil. I think that point holds true but in a different way – what seems to hold true is that the degree of local competition is critical for success.

    The virtualisation of the firm and of business is growing in importance and going to have a stronger influence but it seems surprisingly weak in its influence right now compared to intensity of competition.

  • Gunjan Bhardwaj

    Policy makers dont compete in a virtual world- you are right they compete locally. When the whole world is talking about ‘white spaces’, policy tools cannot remain local in their intent.

  • haydn

    I think there’s this bigger issue you raise. Policy makers swallow the bait on clusters. They are hooked, the get pulled further along and there is no way back. The extent to which clusters have become the only policy option in a world where regional policy is effectively off-limits, that is the problem. It means many policy makers and cluster managers are having to make up the rules and guidelines as they go along, never really sure if it is the appropriate approach to a regional problem of not. This is the worst form of chaos. It is evident throughout Europe and the USA as Governments replaced regional and industrial policy with nothing and forced places to become more active and responsible with no guidelines or skills in economic development of innovation.

  • haydn

    I should have said regional industrial policy

  • Paul Hobcraft

    I do think Clusters have not delivered what they promised and with networks changing the complexity I think it will alter the thinking. Do we really need close proximity? With the infrastructure today we can provide design, supplies and movement of finished goods in a very timely fashion that relies on a supply chain than proximity. Collaboration can be put onto a range of different platforms, it partly depends on the end result need

  • Gunjan Bhardwaj

    Exactly this is what I argue in the third part..on cluster performance. Who defines cluster performance today? We should be asking those for whom these initiatives are designed for- the SMEs and down the road the ROs, MNCs and others..we need an inside out outlook.

  • Jeff Lebow

    I did not see any attention to one of the primary drivers of successful clusters, the ability to attract talent. While information can be accessed globally, people still have to live somewhere and in the “knowledge” age, they will gravitate to regions or subregions that offer multiple employment options, a primary benefit of a geographic cluster.

    It might be possible that the failure of clusters to deliver may have more to do with the global recession and the unsustainability of the capitalist economic model, which depends on continuously increasing consumerism.

  • Gunjan Bhardwaj

    I believe this has changed, now firms drive talent stronger than regions. People are also more mobile.

    Regarding failure of clusters, I believe, the first is obvious- the downturn impacted all firms and clusters arent aloof but this cant be the only cause.

  • Marco Conte

    I would not say that the cluster model is obsolete as such, it is a necessary condition, but no more sufficient.
    Old style services provided via the clusters are obsolete (such as incubation services and so on).
    New style services are required, addressing demand creation, support to the creation of new ventures with international attitude and access to new markets/competencies.
    Support to explore Open Innovation is also required.

  • Pingback: Vers une nouvelle génération de clusters d’entreprise ouverts, associant réseaux sociaux, open innovation et coworking ? : Entreprise Globale

  • via fCh

    “The fact remains however, that in the age where globalisation and technology has virtually made spatial concentration irrelevant,” unless this a rhetorical device to pull the reader in, I find it a very problematic statement. What does the author support his thesis on? The age old concept of ‘structural holes.’ I write old age, for we’ve had clusters throughout time, and they tend to stick to the initial geography, unless big disruptions take place… Structural holes, while potentially helpful to make a point, are misplaced here.

  • Gunjan Bhardwaj

    You can refer to articles published by Ghoshal, Prahlad,Govindrajan or for an easy and interesting reading- The world is flat by Thomas L Friedman. There you will find a myriad of examples to prove the point- that spacial concentration is slowly becoming irrelevant because of technology and globalisation.

  • Udo J. Mannes

    EUREKA Clusters
    (see )

    correspond, as I see it, your 3 points, as they are not geographical oriented, well tracked and quite succesfull. The EUREKA clusters seem to constitute an example for dynamic cluster development according to your ideas.