Recent Research: Manufacturing Productivity Varies by Sector; Knowledge Spillovers Bounded by Distance

As community leaders plan the physical development of their regions, some recent research may offer insight into the benefits of encouraging close proximity between firms. A group from Statistics Canada has published a paper exploring the various gains in productivity that manufacturing firms experience due to geographic concentration. In Urban Economies and Productivity, John Baldwin, Desmond Beckstead, W. Mark Brown and David Rigby use longitudinal data from a collection of approximately 20,000 manufacturing establishments across Canada.

 

Baldwin et al. incorporate theories that originated a century ago from the economist Alfred Marshall. According to Marshall, gains in productivity are due not only to the operational activities that occur within a firm, but also to a firm’s physical location. In his work, he identified three explanations how the geographic clustering of firms can increase their performance. First, collocating firms produce a larger pool of skilled labor with specialized skills. Second, information is transferred between firms in close proximity, resulting in knowledge spillovers. Third, clustering encourages the development of upstream industries that may supply components such as materials and equipment.

 

By using information from the Canadian Annual Survey of Manufacturers to produce measures of these factors, the authors found that for the entire collection of manufacturers across sectors, productivity performance is positively influenced by all three of Marshall’s aforementioned factors.

 

The authors then divided the manufacturing firms into five sectors, defined by the circumstances that influence their competitive strategies. These circumstances were: access to natural resources, labor costs, scale economies, product differentiation, and the application of scientific knowledge. They found that the causes of increased productivity varied from sector to sector, with no single factor statistically significant across all five sectors. For example, labor market pooling had a significant effect on productivity for natural resource-based industries, labor-intensive industries, and product-differentiated industries.

 

Additionally, the authors examined the influence of knowledge spillovers within a determined geographical distance. They used the number of manufacturing plants as their measure for the strength of knowledge spillovers and included two variables of this type in their regressions — one for the number of firms within a 10-kilometer radius (about 6.2 miles) and another for the number of firms located 10-50 kilometers from each other. They found that firm productivity increases and is significant when the number of plants from a similar industry within a 10-kilometer radius increases, but sees little impact among firms exceeding 10-kilometer plant density. Thus, they claim, the impact of knowledge spillovers is spatially dependent.

 

Urban Economies and Productivity is part of the Economic Analysis Research Paper Series from Statistics Canada and is available at: http://www.statcan.ca/english/research/11F0027MIE/11F0027MIE2007045.pdf