Andrew

View from the debate over clustering

Hubs, clusters, technopoles – whatever you call them, these special places are supposed to foster intense collaboration. But are there even more effective collaborative environments, and how would you recognise one?

By Prof Philip Cooke, Centre for Advanced Studies, Cardiff University, and Development Studies, University of Aalborg

Professor Philip Cooke is a research professor in regional economic development and founding director of the Centre for Advanced Studies, University of Wales, Cardiff. His research interests include studies of biotechnology, regional innovation systems, knowledge economies, entrepreneurship, clusters and networks. He advises the UK government on innovation. He also advises regional and national governments, the EU, OECD, World Bank and the United Nations Industrial Development Organization on regional innovation systems.

There's a great deal of interest in using geographic clusters to increase industrial collaboration and hence innovation. The idea has its strengths and weaknesses: an article in the 11 October 2007 edition of The Economist, entitled 'The fading lustre of clusters', criticised clusters from Cambridge to Singapore. I hope to present a more level-headed perspective on their usefulness, and demonstrate that clusters can help counterbalance the management and research weaknesses of large firms. I will also argue that clusters may be an important step on the road to a less specialised, more balanced form of entrepreneurship that is better at promoting the spill-over of knowledge and the cross-fertilisation of innovation than simple clustering.

The idea that the best way to promote the spill-over of knowledge between organisations is to put them near each other has long overtaken the idea of relying on their similarities to do so. Why? I refer readers to their industrial geography schoolbooks. Remember how we learned industry was organised at the beginning of the Industrial Age? British economist Alfred Marshall characterised it superbly as being based in 'industrial districts'. Think of Manchester for cotton, Limoges and Stoke-on-Trent for crockery, Sheffield and Remscheid for cutlery and Detroit for autos. Kortrijk and Kidderminster had more than their initials in common: both specialised (and still do) in making carpets.

Macclesfield was once Britain's silk capital, specialising in flags and white mufflers for sailors, and taffeta with golden threads for captains' wives. Now its silk connections only exist in industrial museums. However, go to Como in Italy and see a modern European silk capital, the source of many a businessman's luxury neckwear. Como survives and thrives against all the odds of its small-scale production 'inefficiencies'. What keeps Como going? Its design-intensive firms using state-of-the-art design, cutting and finishing technologies.

Como is one of innumerable modern Italian industrial districts that fascinate cluster-watchers everywhere. In the 1970s, these north-central Italian districts were shown to out-perform Italian export norms and equivalent businesses located elsewhere. They prospered, increasing the number they employed from perhaps 200,000 people in the 1950s to more than two million by the early 2000s. Most of the world's spectacle frames and ski boots come from Montebelluna and Belluno. Biomedical instruments concentrate in Mirandola, while high-grade knitwear is found in Carpi, near Bologna, as are most of Italy's ceramics production and exports at nearby Sassuolo. The list is almost endless and the 'Made in Italy' branding is testimony to the high quality of the goods exported. It was this unexpected phenomenon of collaborative design and innovation in small and medium sized firms, enabled by proximity, which stimulated interest in what came to be known as 'clusters'.

The Silicon Valley phenomenon of the 1980s appeared to have similarities with those Italian districts. The Valley's microchip firms were small or medium sized, undertook collaborative innovation and accessed local risk capital. One of the Valley's first silicon chip engineers turned away from semiconductor production to venture capitalism. Gene Kleiner founded the Valley's most successful investor, Kleiner Perkins Caufield & Byers. This firm eventually developed a keiretsu system, mimicking the way Japanese industry formed itself into mutual support groups by outsourcing work among the more than 200 firms in which it had invested.

Venture capitalists like 'hands-on' relationships with their investments. Being within an hour's drive of Kleiner Perkins' Sand Hill Road, Palo Alto, headquarters meant these firms formed a cluster that enhanced the all-important knowledge spill-over among its members. They developed synergies through an evolving 'community of practice', as John Seeley Brown and Paul Duguid, veterans of Xerox's Palo Alto Research Centre (PARC), put it. Eventually Xerox PARC engineers preferred to outsource business to other Valley firms rather than interact with engineers elsewhere within their own company. Why? Because the PARC researchers had taken on the Valley's engineering culture – which emphasised 'can-do' attitudes, non-hierarchical management and, above all, innovation. This culture formed the basis of the Silicon Valley phenomenon, attracting large firms to seek solutions from Valley firms and, eventually, to open 'observatory' laboratories there.

In the mid-1970s, a biotechnology industry was born when Genentech began engineering human insulin in south San Francisco. The genetic engineers working in the pioneering firms that span out from academic research labs developed a culture similar to that of the Valley's electronics engineers. At the same time, even as observers worried about the decline of the mini-computer industry clustered around Boston's Route 128 (in part because of the rise of networked personal computers championed in the Valley), Harvard and MIT were spawning their own biotechnology cluster.

As the bio-economy has evolved, some spinouts have grown to corporate scale, even though they operate in very different circumstances from those in information and communications technology (ICT). Testing and trials take more than a decade (unlike in ICT) and finance can dry up with frightening speed. So 'big pharma' is important in determining which companies thrive and which merely survive, by virtue of its investments in promising drug candidates. Big pharma outsources discovery and development work to dedicated biotechnology firms, because its research culture lies in fine chemistry rather than biotechnology. Much like the Sand Hill Road venture capitalists, the big pharma companies like to be close to the action, as shown by the relatively recent arrival in the Boston area of Pfizer, Abbott, Amgen, Novartis and Wyeth. This is why Boston (and Cambridge, MA) has become big pharma's most important bioeconomy cluster. (I expand on this in my 2007 book Growth Cultures: The global bioeconomy and its bioregions)

Many firms that join clusters are satisfied with their discreet charm. But some are not. Most recently, the convergence of biotechnology, ICT and nanotechnology has given rise to 'green clusters' hosting 'green innovation' – currently the fastest growing high-technology industry. Although companies and research organisations are physically gathering to form these clusters, they prefer to think of what they are doing as creating a 'platform', since the convergence they seek involves not only their technology but also knowledge drawn from industries such as agro-food, energy, transportation and recycling. Their innovation involves far more lateral collaboration than the vertically oriented innovation that tends to happen in today's specialised clusters.

The power of these platforms is only limited by the extent to which knowledge can be usefully cross-fertilised from one context to another. Putting people and facilities into these platform regions means that knowledge can be shared instantaneously by chance meetings and happy accidents, rather than being brokered through conference presentations, emails or search engines. Proximity is the means for accessing knowledge spill-overs whose secrets are, to quote Alfred Marshall, 'in the air'.

What is the advantage of a platform perspective? Let's look at two highly successful platform regions, both of which have attracted large firms to access the knowledge and expertise that each has gathered.

The first platform is based in the Greater Boston area, which hosts healthcare assets such as Harvard Medical School, Massachusetts General Hospital and MIT. It also has the world's leading biotechnology cluster (as discussed), a strong medical instruments industry, a software business and 150 venture capital companies. The area is so strong that since 2002 many of the pharma companies mentioned above have set up 'observatory' labs or even moved global R&D headquarters there, to access the talent, advanced know-how and investment targets in the area.

The second platform is North Jutland, Denmark, home of the wind turbine industry. Of the Danish Wind Industry Association's 70 members, 50 are based in this region between Aalborg and Aarhus, providing a complete supply-chain for the manufacture of wind energy turbines. The cluster is also home to Denmark's leading solar thermal energy businesses. Foreign firms such as Siemens of Germany, Gamesa, Spain's leading producer and Suzlon, India's equivalent, have already moved in to access the local wind turbine talent and know-how. At Aalborg nearby, a wireless telecommunications cluster known as NorCOM has grown up, drawing in Texas Instruments, Siemens, Flextronics, Cambridge Silicon Radio and Motorola. Spillover from NorCOM has already helped the emergence of a biomedical instrumentation cluster, which has now attracted investment from Nycomed and Novo Nordisk.

What's it like to work in one of these places? I got some idea from talking with the head of a GE Medical Technology facility in a specialised bioengineering cluster at Uppsala, Sweden, who is also the head of the cluster association. Did he have conflicts of interest, I asked: what would happen if an innovative firm arose as a competitor to GE's technology? He said that as head of the cluster, he would do everything in his power to nurture the competitor. If its work proved of interest to GE, he would shift to his GE role and try to acquire it. No conflict.

The emergence of platform regions may influence innovation policy to shift away from supporting clusters and sectors, which are silo concepts, towards supporting innovation that crosses boundaries. US science policy has already made that move, supporting research into topics such as complex systems, human-computer interfaces, resource exploitation and homeland security, whose solutions lie across disciplinary and sectoral boundaries. Many OECD member countries are still supporting narrow topics such as ICT, biotechnology and nanotechnology.

If innovation policy does shift towards supporting platforms, it implies greater backing for interdisciplinary teams drawn from industry and academia. This will mean challenges for industrialists, as they try to apply the matrix management techniques they use on internal innovation projects to these more collaborative approaches. The trust vital to making these more collaborative projects work will best be built up by being present in the platform regions, where face-to-face contact and a confirming handshake are vital to creating the social capital that innovation across interfaces requires.

Industrialists cannot stand alone any more. The world is spiky, not flat, and there aren't very many of these platform regions in which knowledge flows freely among a number of world-class disciplines. R&D managers should consider whether they have the best possible environment to exploit their knowledge and that of partners if they are not attached to a relevant spike.

Prof Philip Cooke

Centre for Advanced Studies, Cardiff University, and Development Studies, University of Aalborg

cookepn@Cardiff.ac.uk

Issue 13 Spring 08

View from the debate over clustering