The Categorization and Assessment of Industrial Policy: Automotive Policy in the Semiperiphery

Publication Type:

Conference Paper


Gerpisa colloquium, Paris (2019)


The Categorization and Assessment of Industrial Policy: Automotive Policy in the Semiperiphery

The positive attributes of manufacturing are well-established. It stimulates economic growth (Szirmai 2012; UNIDO 2013), generates exports (Fryges and Wagner 2010; Roper and Love 2002), spurs R&D and innovation (Eurostat 2013; National Center for Science and Engineering Statistics 2016; Clark 2016; Stöllinger 2016; Statistics Canada 2018a), outperforms other industries in terms of productivity (Rowthorn and Coutts 2004; Rodrik 2013) and generates high economic multiplier benefits (Rachwal and Wiedermann 2008; Zhang, Kinnucan, and Gao 2016). It also produces high wage employment (Tregenna 2009; Pierce and Schott 2016), “especially for workers who would otherwise earn the lowest wages” (Helper, Krueger, and Wial 2012, 3). It is for these reasons that the instruments of industrial policy have been deployed to support manufacturing with greater consistency and intensity than any other sector. In the past decade, the automotive manufacturing industry in particularhas generated disproportionate attention from policy-makers and academics. This is an outcome of the need to resuscitate the sector following the recession (Livesey, 2012), the advent of new vehicle and production technologies, and the ongoing recognition of thesubstantial upstream and downstream benefits the industry provides (Zhang et al., 2016; Mordue and Sweeney, 2017a).

Therefore, the indispensable nature of manufacturing in general and automotive manufacturing in particular is well-established. So too, is the importance of industrial policy. What does not exist, however, is a comprehensive, quantifiable means by which to document, categorize and compare industrial policy within jurisdictions, across jurisdictions and over time. Our work attempts to bridge that gap and is inspired by the work of O’Sullivan et al. (2013) who offer a framework to illustrate and compare industrial policy tools viatwo dimensions: factor inputs and intervention levels. However, O’Sullivan et al. do acknowledge shortcomings with their framework. These include challenges associated with gathering reliable information, difficulties recognizing if ‘new’ programs are incremental or merely rebranded versions of previous undertakings, the fact that not all announcements manifest in programs, and capturing programs at the subnational level. It is for these reasons that O’Sullivan et al. do not ultimately deploy their framework for detailed analysis. Thus, it remains a vague and undeveloped model for how industrial policy might be evaluated.

Our methodology addresses shortcomings identified by O’Sullivan et al. To do so, we limit our attention to programs that are applicable to automotive manufacturing (but not necessarily restricted to automotive manufacturing) and build a database of programs at national and subnational levels in targeted jurisdictions at designated points in time. We build from the work of O’Sullivan et al. by clarifying the two dimensions anchoring the framework (“Factor Inputs” like labour, research, capacity etc. and “Intervention Levels” like single establishment, firm, specific/targeted Industry, multiple sectors etc.) and assigning greater precision to the descriptions for each aspect of each dimension. Then, we search, analyze and categorize various programs from disparate sources. It is, as O’Sullivan predicts, a labour-intensive undertaking.
As described, we limit our analysis to automotive-applicable programs. Moreover, we focus our investigation on national and subnational automotive jurisdictions that are linked via three factors: 1) the automotive industry is important (as evidenced by the fact they build more vehicles than are purchased), 2) labour costs are high relative to regional competitors, and 3) production is not anchored by an indigenous OEM. We consider countries like Canada, Austria, Spain, Belgium, and possibly the UK as holding those attributes. While there have been previous attempts to categorize such countries, there are shortcoming. For example, Sturgeon and Florida (2000) discussed “Type 3” peripheries of large existing market areas, grouping Mexico, Canada, Spain, Portugal, and Eastern Europe together and describing their principal strategic attribute as being a “proximate low-cost location from which to supply Type 2s” (the United States, Northern Europe, and Japan). Neither now nor then, has Canada (and possibly Spain) been a low-cost location, particularly in the context of jurisdictions like Mexico or Eastern Europe. Pavlinek (2018) provides a category deemed the “semiperiphery”, and lists the UK and Canada as examples. The attributes he assigns are more accurate, but the category and discussion are tangential and abbreviated.

Our research seeks to expand knowledge of the semiperiphery and in so doing generatethree policy relevant research outcomes. First, itdevelops a methodology to categorize and compare industrial policy between jurisdictions and over time. Second, it modernizes the literature in economic geography related to the role of individual countries and regions within the automotive industry’s global production network. Third, it provides insight into options available to policy-makers in semiperipherycountries like Canada, Austria, Spain and Belgium. Because they are neither low cost members of the “integrated periphery” nor “core” markets hosting indigenous OEMs, they have unique vulnerabilities.




  GIS Gerpisa /
  4 Avenue des Sciences, 91190 Gif-sur-Yvette

Copyright© Gerpisa
Concéption Tommaso Pardi
Administration Juan Sebastian Carbonell, Lorenza MonacoGéry Deffontaines

Powered by Drupal, an open source content management system