Defending Industry Clusters

A few weeks ago, I came out in support of Vivek Wadhwa's critique of industry cluster economic development. My spin on it is pushing people-based initiatives in front of place-based programs. I'm not a fan of placemaking, particularly concerning talent attraction and retention.

As Mike Knutson (Reimagine Rural) once wrote about me, "There’s usually no middle ground with Russell, but he shares innovative ideas – and usually supports them with research."

Once again, I'm taking a strong stand against a policy framework. However, I am aware of examples and research that run counter to my position. The Economist offers one anecdote about the tech cluster in the Bristol (UK) region, "Silicon Gorge". It's a government-driven success story, a cluster that works. I see a different angle:

Inmos was sold to Thorn EMI in 1984, but the firm remained an important training ground for the electronics graduates who flocked there. David May left a teaching post at Warwick University for a job at Inmos in 1978. He helped design the “transputer” technology used in countless set-top boxes. He continued his career at STMicroelectronics, a Franco-Italian firm that bought Inmos in 1989.

Mr May is now a professor of computer science at Bristol University and chief technology officer at Xmos. That firm designs microchips that can be easily programmed by makers of niche goods such as audio equipment, musical instruments and fitness monitors. It started as a final-year student project and now employs 50 people, mostly in Bristol.

Mr May has also worked as a technical adviser for other start-ups, including Picochip, which is based in nearby Bath, a city famous for its Roman spa. “We are the only chip company in a world heritage site,” says the firm’s boss, Nigel Toon. Picochip is the market leader in femtocells, which redirect mobile-phone signals through landlines, boosting coverage in remote areas and prolonging battery life. By economising on scarce mobile spectrum, femtocells might prove most useful in managing the rapid growth of internet and video traffic to smartphones and tablet computers.

Before he joined Picochip, in 2002 Mr Toon was one of four co-founders of Icera, which makes chips for mobile devices and employs more than 300 people, mostly in Bristol. The twist with Icera’s chips, says Steve Allpress, a co-founder, is the efficient way they “speak” to base stations, allowing for quicker and cheaper data downloads. The two other co-founders, Stan Boland and Simon Knowles, previously carved a successful venture out of Acorn, a British computer firm that also sired ARM.

That's a network of talent, not an industry cluster. The distinction is semantic. Mr May and his connections would start a cluster wherever they were located, like research universities competing for the top professors in a given field. Yes, the talent is clustering in Bristol and Silicon Valley. That's still not the same thing as clustering biotech firms in a suburban park. People develop, not places.