Globalization hasn't made manufacturing clusters obsolete, but the geographically concentrated pockets of industry have to be smart to ensure their survival, according to new research from Harvard Business School.
Gary P. Pisano, the Harry E. Figgie, Jr. Professor of Business Administration at HBS, and researcher Giulio Buciuni, of the University of Venice Ca' Foscari, address the question of when clusters survive and when they fail in their May 2015 working paper, Can Marshall's Clusters Survive Globalization?
“I think people in general think well, there's globalization so there's no need for clusters”
Pisano and Buciuni looked to four industrial clusters in northeastern Italy for their answer. Italy "is a great laboratory because it's been historically organized around these very specific districts all over the country. Every area is associated with a particular industry," Pisano says.
Clusters are not a new concept, notably studied in the United Kingdom by Alfred Marshall—he called them industrial districts—in the early 1900s. Manufacturing clusters can seemingly happen in any industry in any location, from winemaking in Northern California to automaking in Detroit. Clusters typically build up around a geographic location where natural resources, an appropriately educated labor force, and a university or other research institution co-mingle.
In recent years, some economists have argued that manufacturing clusters are dying out because geographic location is less important to business success. In today's global markets, companies have many choices to procure what they need to develop, build, and sell product.
Pisano and Buciuni wanted to test the theory of the dying cluster. "I think people in general think well, there's globalization so there's no need for clusters," Pisano says. "What we came to is it really depends. It's not all or one."
The four clusters used in the field study had differing fates. Two—sports shoes and wooden chairs—declined while the other two—high-end women's shoes and furniture—not only survived but thrived.
The two shoe clusters studied are physically only 50 kilometers apart, but worlds apart in their success. The sports shoe cluster in Montebelluna dates back to the 1800s when local artisans began making hiking boots. The cluster later expanded into ski boots and sneakers, and drew investments from major brands like Nordica and Nike.
Lead firms started offshoring production to cheaper economies like Romania and China, separating the R&D from production, which was high volume and mostly unskilled. They created molds for soles and other components that then could be shipped elsewhere for production. As a result, the local workforce dropped by 15 percent between 2006 and 2012.
"What happens in some contexts is that you no longer have to be close to the sources of production. That happened with the ski boots. Once you've made the mold, that boot can be produced anywhere in the world," Pisano says.
The women's shoe cluster also sprouted in the 1800s, when leather shoe production began in Riviera del Brenta. That cluster maintained production locally while also investing in R&D and engineering. The region is now producing an estimated 90 percent of women's luxury shoes that sell for $500 or more per pair.
Why Clusters Still Work
The cluster's success is largely due to a lead firm acting as what Pisano calls a knowledge integrator. The knowledge integrator is deeply embedded in the local network while also having access to global markets. It bridges the two worlds, helping the cluster maintain its competitive advantage. In one Brenta firm, skilled workers translate international designers' sketches into prototypes, which pushes local suppliers to advance their production capabilities continuously.
The other two clusters studied—wooden chairs in Manzano and furniture in Livenza—show a similar pattern. Taking root in the 1800s due to the proximity of raw materials, both clusters boomed from the 1960s through the 1990s. Manzano went from producing a third of all chairs sold in the world in the 1990s to hardly any production today. The workforce dropped by 44 percent from 2006 to 2012. One firm has succeeded by moving downstream, establishing a global network of retail stores and showrooms, but the rest have petered out.
"That district got clobbered," Pisano says. "They were making wooden chairs—nice wooden chairs—but the market was shifting to plastics, composites, different designs, et cetera, and they weren't changing their capabilities."
Sixty kilometers away, Livenza is thriving with its furniture production. The difference, Pisano says, is again the presence of a lead firm acting as a knowledge integrator. In this case, the lead company is IKEA's largest European supplier. The firm developed a way to make high-quality curved laminates for the Swedish retailer, and continues to come up with new processes.
"If you're a big furniture company and you're investing in more advanced equipment, that's going to push your local suppliers into having to innovate," Pisano says. "If you want to work with us as a local supplier, you have to be able to achieve this level of quality or this level of flexibility."
Pisano stresses the importance of staying ahead of the market. That has helped the lead firm in this case grow from sales of 20 million euros in 1997 to 500 million euros today.
"They drove the innovation. They weren't waiting around to be told," Pisano says.
Another finding of the research is that manufacturing can be what Pisano calls "sticky."
"A lot of the thinking today on manufacturing is as wages change, manufacturing moves. And you'll get these very optimistic reports about the return of manufacturing to the United States from places like China and Taiwan, because wages will become high enough there that we become more efficient," Pisano says. "But that's assuming manufacturing is very mobile, and that's not always the case."
Sometimes manufacturing won't move because the process requires suppliers or specialized skills already located in an area. Pisano cites the bicycle industry as an example. When the industry moved from steel tubes to carbon fiber ones, the cluster took hold in Taiwan because that's where the expertise is in fabricating and molding carbon fiber.
"If the prices go up there, it's not like you can just reestablish that capability locally," he says.
Trade agreements opened the world economy, and factor cost differentials, like labor costs, became more important, Pisano says. That's why there was a huge outflow of unskilled apparel manufacturing from the United States. "There's nothing that makes it sticky," he says.
Can Government Create A Cluster?
Pisano is skeptical of efforts by the federal government to create new clusters in the United States. He points to the rise of biotech in Harvard's backyard, which began despite government opposition in the 1970s.
Clusters "tend to emerge much more organically, and they typically emerge organically with a lot of competition between players," Pisano says. Government-created initiatives engender cooperation rather than competition. "All of these clusters, even in Italy, are very, very competitive."
Executives whose firms are already in a cluster need to ask themselves if the cluster is vibrant, and what their firms can do to ensure future success, Pisano advises.
"Clusters that aren't dynamic, that don't innovate, die today," Pisano said. "That's one of the challenges of globalization. It used to be that clusters were very static."
If there isn't innovation, newer and more entrepreneurial firms will locate elsewhere and create a new cluster. Once the cluster erodes, it's hard to rejuvenate it.
"When they actually start to decline, they come apart pretty quickly," Pisano says.
When making supply chain decisions, executives need to understand it's more nuanced than global versus local. "In some cases global means everything is scattered, and nothing has to be close to each other. But in others you are global in terms of supplying a global market but from a clustered location," Pisano says.
Global supply chain is a meaningless term, Pisano says, because they can be configured in many different ways. "There are different choices that emerge."