GLOBAL: What makes science parks successful?

Clusters of innovative firms are an old phenomenon. In the UK during the first Industrial Revolution, for example, the cotton industry was heavily concentrated in Lancashire within the Oldham-Bolton-Manchester triangle. In more recent times, clusters of high-tech firms, prime among which is Silicon Valley in California, have gained fame and are routinely referred to as role models for promoting innovation, successful commercialisation of research and economic growth. What is it that makes clusters so attractive?

here is extensive empirical evidence for thinking that clusters generate some tangible benefits, such as knowledge spillovers, the sharing of inputs and forward and backward linkages to research innovation, which make firms within the cluster more productive and innovative. Some firms might even never have been founded outside of such clusters.

While there is robust empirical evidence on the positive effect of clusters on firms' (innovative) performance, far less is known about the processes that give rise to clusters.

Natural advantages may play a role, such as the local availability of raw materials, proximity to the sea, or just a pleasant climate that makes it easier and cheaper to attract qualified labour. There is also some evidence that the agglomeration of high-tech firms occurred naturally in some places, such as Boston's Route 128, from a combination of the availability of physical resources, such as high-skilled labour, and world-class research institutions.

Spin-offs from large existing companies and universities appear to play a particularly important role in kick-starting the clustering of innovative companies. However, this process is far from automatic. For example, evidence from the UK suggests that it is low-tech manufacturing industries, such as textiles and cutlery, rather than high-tech industries, that are geographically concentrated.

Is there a way to proactively create high-tech clusters?

Science parks represent a popular policy intervention to promote the clustering of high-tech industries. Policy-makers around the world follow the deceptively simple 'science park formula' to create high-tech hubs: pick your preferred location, preferably in geographical proximity to a university, build some real estate to accommodate companies, offer financial support in the form of subsidies or tax breaks and the best entrepreneurs will come and innovation and its successful commercialisation will happen.

Examples of this seemingly simple approach to creating high-tech clusters abound around the world.

China set up its first park, the Beijing Zhongguancun Science Park, as early as 1988. Its proximity to top Chinese universities and research institutes of the Chinese Academy of Sciences had a major impact on its success. Zhongguancun was a motor of the Chinese semiconductor, computer and telecommunication industry during the 1990s, giving birth to such world-leading companies as Lenovo and the Founder Group.

In February 2010, Russia announced a lighthouse project that aims to build the Innograd science park in close geographical proximity to the recently established top-notch Russian business school in Skolkovo near Moscow. To ensure the project's success, the Russian government has pledged generous public financial support and tax breaks and made substantial efforts to attract large foreign high-tech companies to set up research activities in the park.

Even more recently, in November 2010, the UK government launched a similarly ambitious programme, East End Tech City, which aims to exploit real estate and facilities constructed for the 2012 Olympics to turn London's East End into a global technology hub comparable to Silicon Valley.

While the Chinese park appears to have been successful during the 1990s in aiding the emergence of the Chinese electronics industry, does this imply that innovation can be created and successfully commercialised through this simple 'science park formula'? Will Skolkovo or the East End Tech City be successful? How should they be designed to be successful?

Although science parks are commonly believed to generate 'by construction' ample benefits for their tenant firms, there is surprisingly little quantitative evidence supporting this claim and even less is known with regard to the determinants of these benefits.

Probably the most fundamental question to ask in this context is what the ideal composition of firms in a science park should be in order to maximise the benefits of clusters. Should parks host a set of homogeneous firms, for example only bio-tech, or are they better served by accommodating a set of heterogeneous firms, say a mix of bio-tech and IT companies?

Using data on two geographically adjacent science parks in the UK, the Cambridge Science Park and St John's Innovation Centre, recent research suggests that firms within the same industry benefit most from mutual knowledge spillovers. The evidence on science park-wide spillovers across firms from different sectors is, at best, weak.

If the benefits of science parks occur mostly through face-to-face contact between entrepreneurs and employees, an intuitive explanation for the result is that, for example, a software engineer finds it easier and more insightful to exchange knowledge with another software engineer from a different company than to exchange ideas with a biochemist.

Even moving between different employers in a cluster and transferring intangible knowledge in this way is easier for employees if firms operate in similar areas. The data on the two UK science parks also reveal substantial churning of companies; that is, entry into and exit from the parks. Therefore, even when they are in the park, firms are far from certain to remain there. This continuous change creates complicated dynamics in the interaction of companies in a park.

These findings imply that the selection of firms into science parks may have an important impact on the success of the park in fostering firms' performance and innovation. Moreover, maintaining the right balance of companies in a park over time is difficult due to the continuous changes caused by growth, failure and the moving in and out of companies.

Having a clear strategy at the time of establishing a park and focusing on certain types of firms carrying out similar activities that fit the research profile of the university linked to the park may improve performance.

This suggests that active management of science parks and targeted selection of tenant firms has an important role to play in ensuring firms benefit from clusters. Simply offering real estate and financial incentives is unlikely to attract and maintain the right mix of companies that maximises the benefits for which science parks are constructed in the first place.

* Christian Helmers is a research economist in the spatial economics research centre at the Centre for Economic Performance, London School of Economics, and also a researcher in the economics of business department of the Universidad Carlos III de Madrid. He presented a paper on his research at the Royal Economic Society's annual conference in April.