Mark Alan Hughes

Impact of innovation clusters

Distinguished Senior Fellow, PennDesign; Senior Fellow of The Wharton School’s Initiative for Global Environmental Leadership, Aug 10, 12.

Tagged with economy, government, cities

Much like watching your country’s flag rising at an Olympic medal ceremony, it is exciting to see your city in the running for the mantle of “innovation cluster.”

The term is a mantra of cities these days, much as “green jobs” and the “creative class” stood at the buzzy center of previous idea cycles. Indeed, these conceptual bumper stickers have much in common. But the “innovation cluster” has the firmest intellectual foundation, being part of a conversation about cities that has continued vigorously for over 100 years.

Alfred Marshall was a British economist who wrote the very influential 1890 book Principles of Economics. Like Shakespeare, Marshall is a writer you have probably quoted without ever knowing it: “On the margin,” “economies of scale,” “supply and demand curves” and “diminishing returns” all first appear in their modern form in Marshall’s book. He was also the first to discuss the ways in which cities improve the performance of markets, burdening a cool idea with a nerdy word: “agglomeration.” Agglomeration is when a firm is more productive because of increased activities outside the firm itself and beyond its control.

What is the difference between a firm being located in Philadelphia or in the middle of Pennsylvania? Economists and geographers have described many differences between the two locations. But one idea generates the most attention these days — the notion that a company located in dense Philadelphia may be closer to other companies in its own industry than it would be if it was located in a less dense locale. This is a called a localization effect, and when localization is strong enough the resulting concentration within an industry is called a cluster. Famous examples are software in the Silicon Valley of San Jose, Calif. and its suburbs, carpets in Dalton, Ga., and furniture in High Point, N.C. The idea is that the bigger the cluster, the more productive the firm located in that cluster.

Why would being in a cluster cause a firm to be more productive? It was Marshall who first provided three of the strongest explanations, which are still being debated and measured by scholars like Michael Porter and Ed Glaeser today.

The first explanation is called input sharing. In volatile knowledge-intensive activities, there are more firms involved in the making of something (vertical disintegration, for you B-schoolers) in order to manage the risk of needing some specialized input. The classic example is movies: Why would every studio have its own CGI department when they can always purchase the latest technology from the company Industrial Light & Magic? In a cluster, a firm has access to more input sharing and thereby lowers its purchased input costs.

The second explanation of localization effects is called labor market pooling. Here the idea is that larger clusters offer better matches between workers and firms. When quits and fires happen, it is easier for a worker to get a new job, and for the firm to fill the old job in a location where an industry and its specific skills are more concentrated into a cluster.

The third explanation is called knowledge spillover. Again, Marshall got there first: “The mysteries of the trade become no mysteries; but are as it were in the air…” (The literary skills of economists have declined precipitously over the past century.) This idea is the most nuanced and, therefore, the hardest to measure. But it also seems like the big enchilada, especially when it comes to intentional change designed to encourage cluster development in particular places. Do ideas leading to innovations and entrepreneurs leading to growth benefit from clusters — from the concentration of industry-specific activities at a geographic scale, allowing for face-to-face interactions? If the answer is yes, and the evidence is mixed, then the implications are huge.

First, knowledge spillovers depend heavily on the type and volume of knowledge produced locally. You can import some of it, but that’s far less efficient than exposure to new knowledge before anyone knows it’s worth importing. Network connectivity among existing local knowledge also creates opportunities for idea brokerage in place, increasing knowledge spillover.

Second, communications technology doesn’t undermine the importance of concentration. Sure, anything that can be written down can be conveyed over any distance between two internet connections. But there is more than just codified, technical knowledge in these spillovers. The so-called tacit knowledge of individuals is where much of the innovative spillover occurs, and this is far more difficult to convey over distance.

Third, even within a local concentration, cognitive and social distance will matter as much as physical distance to a cluster: Knowledge has to spill over just the right distance to work. If firms and people operate with very different knowledge, languages, organizational cultures and settings, then knowledge can’t spillover — universities and corporations, psychologists and engineers, elected officials and asset managers, all have plenty of sources for miscommunicating. But if there is too little cognitive distance, then there will be competitive reasons to not communicate and impede knowledge spillover.

All of this suggests that there are compelling reasons to identify, foster and strengthen innovation clusters as a strategy for regional prosperity. Building on the work of Harvard’s Michael Porter, the non-profit think tank Council on Competitiveness and the Brookings Institution have led the charge on this issue, with the Institution’s own Bruce Katz and Jennifer Bradley making the case in a recent Forefront article.

...