Industry And The Urge To Cluster

What drives industry to locate in one region and not in the next?

Economic geography – the distribution of economic activity over physical space – has always been central to economic development. Policy-makers trying to encourage economic activity to locate in under-developed regions want answers: Is it infrastructure? Fiscal incentives? Good business environment? Or could it be agglomeration – the compounding effect of industry clustering in a particular location?

And if the key factor is indeed this critical mass, can the effect run from one type of industry to another? Do existing, more traditional manufacturing clusters attract newer services industry?

The question of where and how services firms decide to locate themselves has become exceedingly central to understanding economic growth and development. Services, and especially knowledge-based services, now account for a greater proportion of advanced-country GDPs, and increasingly so for emerging economies.

New Economic Geography (NEG) theory would argue that agglomeration advantages lock business activity into core regions. The core also supports the existence of intermediate industry in the periphery, and so specialized input-suppliers co-locate close by. For instance, think of Detroit’s production of automobiles and the auto-parts manufacturers who locate in geographically proximate Michigan, Ohio and Indiana.

The theoretical business-economics literature would also argue that manufacturing and services are intricately linked in the production chain. For example, marketing services add the finishing touches in the final stages of a manufacturing process, or research and development services result in increased production within the "real" economy. Service inputs into production, such as design, technological refinements, and branding, account for a major part of value added in manufacturing industries. The result is that it is becoming difficult to identify where the product ends and where the service begins.

These theories have been challenged by claims that services, as compared to manufacturing, are liberated from the tyranny of space, owing to advances in information and communication technologies. In addition, the ability to splice the service production chain more thinly, goes the argument, means that proximity may cease to be an important factor with regard to these industries.

But some empirical research suggests that agglomeration forces may actually be stronger in the case of services – that service industries actually tend to cluster more strongly and more closely to existing urban or manufacturing agglomerations.

How do we know this? It is true that while the interest in urban, regional and spatial economic theory has grown dramatically in the last few decades, empirical research has followed in fits and starts. Some historical evidence shows that manufacturing does indeed precede services, specifically producer services, in a city or city-region. Research in the United States in the mid-1990s, and then more recently, also demonstrates that financial and professional services firms often chose to locate themselves in geographic proximity to established manufacturing industrial areas.

More macro-level North-South models of development also seem to lend credence to the idea that services cluster close to existing manufacturing companies. Research on firm location in Sweden has shown that producer services locate themselves close to manufacturing industry to benefit from accessibility to their customers, but that many producer services also look to supply other service industries. Similar research in Denmark shows that manufacturing and services can be so intricately linked in their production chains that firms across both types may decide simultaneously to choose one location over another.

And there is yet another possibility. Research in Japan’s urban areas revealed that the presence of a large and growing service sector in an existing urban cluster could lead to the displacement of manufacturing units.

There are two basic causes of clustering. The first is regional endowments such as land, climate, and waterways. The second is circularity in location choice, implying that firms want to be where large markets are, and large markets are where many firms are located.

This logic may seem obvious now, but, as economist Paul Krugman notes, that wasn’t really the case before 1991. In the latter half of the 19th century, the emergence of the manufacturing belt in the United States was a turning point in the economic geography of the country. The belt – mainly New England, Middle Atlantic and east-North-Central regions – contained the majority of manufacturing employment up until the first half of the 20th century. It was a classic example of how concentration of firms in one region increased local demand and thus made the area attractive for other firms. Services industries, catering to both final consumption and to manufacturing, soon followed.

What does all this tell us about how governments should focus their energies? If the whole point of policy were to encourage industrial growth in regions that were not previously favored by economic activity, then a multitude of factors would need to be considered: investments in educational infrastructure, and training and development of skilled labor are just some examples.

If policy was aimed at the development of producer services industries, then it seems that a healthy manufacturing sector is vital to a healthy services sector. There is, however, an ongoing blurring of the distinction between what constitutes manufacturing and what constitutes services, and this transformation has stimulated new support functions that feed the production processes of both.

If so, and if the different types of industry do simultaneously co-locate, then...the discussion is akin to going round the Mulberry bush.

Megha Mukim is currently reading for a Ph.D at the London School of Economics. Prior to this she was a visiting fellow at the MacMillan Center for International and Area Studies at Yale University.