Roepke Lecture in Economic Geography—Rethinking Regional Path Dependence: Beyond Lock-in to Evolution

Abstract

This article argues that in its “canonical” form, the path dependence model, with its core concept of lock-in, affords a restrictive and narrowly applicable account of regional and local industrial evolution, an account moreover that is tied to problematic underpinnings based on equilibrist thinking. As such, the canonical path dependence model actually stresses continuity rather than change. The article explores recent developments in political science, in which there have been active attempts to rethink the application of path dependence to the evolution of institutions so as to emphasize change rather than continuity. These developments are used to argue for a rethinking of path dependence ideas in economic geography.

Keywords:

• path dependence
• lock-in
• equilibrium
• evolution
• layering
• conversion
• adaptation

Acknowledgments

This is a revised version of the 2009 Roepke Lecture I was honored to deliver at the annual meeting of the Association of American Geographers (AAG), Las Vegas, in March 2009. I am indebted to Yuko Aoyama for inviting me to give this lecture on behalf of the Economic Geography Specialty Group of the AAG, the Economic Geography journal, and the University of Illinois. I am also grateful to Professor Amy Glasmeier for acting as discussant and to those who also made perceptive and thoughtful comments on the lecture, especially Allen Scott, Peter Maskell, and Jürgen Essletzbichler. In addition, four reviewers made valuable and constructive comments on the original draft. The article is, I believe, much improved as a result of these various contributions. Likewise, I have benefited enormously from discussions with colleagues Ron Boschma and Peter Sunley as part of an ongoing dialogue on the subject of evolutionary economic geography. Of course, any errors of interpretation and argument remain my sole responsibility. The research for this article was undertaken while I held a Leverhulme Major Research Fellowship.

Notes

¹ When I was invited to give the Roepke lecture, I was urged to speak to a conceptual issue and to be provocative. This is what I have tried to do. If this article stimulates debate, regardless of the nature of the reaction, then at least that aim will have been achieved. The danger, of course, is that such an article can provoke varied, divergent, and even critical responses: it can become the victim of its own contentiousness.

² As such, the path dependent process in (1) can be contrasted with the conventional equilibrium process of mainstream economics, which can be represented as where F_x^e generates a unique equilibrium state or outcome, regardless of where the system starts from or which adjustment path is followed.

³ Arthur distinguishes between two kinds of industrial location processes: the case of “pure necessity,” in which the location of a new industry is tied to particular and unique input needs (such as raw materials) that are available only in certain locations, so that the long-run outcome is predetermined, and the path dependence case, in which the long-run locational pattern of an industry depends on “historical accident”—the “chance” location of the initial firms—plus the emergence of agglomeration economies.

⁴ This is not to deny that occasional major shocks can and do occur (witness the current global financial crisis and associated recession) or that such disturbances can produce a “gale of creative destruction.” It is rather to highlight the fact that change is not confined to such “critical junctures” but is an incessant process, as Schumpeter (1942) himself stressed.

⁵Page (2006) presents a useful account of Polya and other probability models of path dependence and shows how such models characteristically converge to a stable, “equilibrium” outcome. Although he acknowledges that “outcome dependence” need not imply a convergence to any equilibrium state, the whole thrust and focus of his discussion is on path dependence models that do lead to equilibrium outcomes.

⁶ I say “ironically” because David has always been at pains to present his path dependence theory as a counter to mainstream economics. On the other hand, one wonders whether by calling it “path dependent equilibrium economics” and by invoking the device of multiple equilibria, his purpose has been to make it easier for mainstream economists to accept the role of “history” and thereby to gain disciplinary acceptance of the notion of path dependence.

⁷ Although even some of these technologies actually evolve and do not remain stable. For example, the Internet—arguably a general-purpose technology—has not become locked into a stable equilibrium state but can be shown to have evolved more or less continually since its initial development (see Boas 2007).

⁸ This is precisely what coevolution means in a regional or local economic context. Of course, some of the coevolving components and externalities need not be local in form or constitution but may involve relationships, institutions, and organizations at the national or even international scale.

⁹ In what follows, I draw on Ph.D. research conducted by Maria Corte-Real (2008). I am grateful to her for allowing me to use her work in this way.