An evolutionary view of the infant-industry argument

Abstract

As opposed to the conventional view of the infant-industry argument as a rationale for temporary trade-impairment measures, this paper submits an evolutionary view of the argument, founded on insights from catch-up experience, as a long-run collective-learning enterprise providing a rationale for trade creation. This is operationalised by introducing system, institutional and policy learning, in addition to firm-specific learning, to complete a triad that unfolds within a multi-phased dynamic sequence eventually leading to infant-industry emergence. We factor-in radical uncertainty, complexity and path dependence, while spelling out positive and normative implications. Crucial interphase transition junctures are identified with emphasis on generative, as opposed to functional, properties.

Keywords:

• infant industry
• institutional learning
• policy learning
• emergence
• dynamics

Acknowledgements

The authors want to thank Richard Nelson and two anonymous referees for their comments on an earlier draft. The usual caveat applies.

Notes

For a statement of this view, see Panagariya (2011).

In his ‘Report of Manufactures’ Hamilton indirectly highlights the importance of learning in various places. Later contributors were List, Raymond, Mill and Bastable. List argued that the defence of unfettered free trade was largely a rationalisation of the interests of incumbent powers to protect their leading positions.

Rodrik (2009) highlights the importance of learning promotion policies vis-à-vis trade regimes in East Asia's catch-up. Taiwan's leading II sector (electronics) has not drawn on tariff protection (Mathews and Cho 2000).

For an early contribution along these lines, see also Cooper and Sercovich (1971).

Some functions-related typologies (Dahlman, Ross-Larson, and Westphal 1987; Lall 1990) did seek insights on policy implications. Performative dimensions such as functions can only implicitly capture discontinuities in capability formation and knowledge flows and their cognitive underpinnings by leaving out the cognitive, ostensive and art–factual dimensions, which are also needed in order to get a comprehensive grasp of the dynamic at play (Becker 2011). In addition, the causal mechanisms involved in firm-specific learning processes and the way such mechanisms relate to (and co-evolve with) policy interventions are still largely unaccounted by functional approaches. An evolutionary account of the II argument cannot fail to address either of these key issues. On length of learning periods, see Bell (2006).

On the concept of behavioural additionality, see Georghiou (2007) and Gök (2010).

Both sectors have been shaken up along the way by major changes in technological trajectory at world level (the advent of recombinant biotechnology and genetic engineering in the early 1980s and that of the Californian ‘new world wine’ approach to wine-making in the 1970s).

We adopt here a broad definition of institutions as comprising habits, routines, norms of behaviour and social conventions, i.e. systems of rules that shape social cognition, actions and interactions of all system actors (North 1990; Knight 2009; Hodgson 2006; Malerba and Nelson 2012). Institutional learning is coterminous with the co-evolution of the system agents, including firms, other sector agents and policy-making agencies.

It is worth noting that institutional learning cuts across the market/non-market divide.

On Taiwan, see Wade (1990), Hong (1997), Mathews and Cho (2000), Amsden and Chu (2003) and Breznitz (2011). A two-phase approach is often resorted to in this literature regarding the learning process in semiconductors. First, ITRI/ERSO acquire a foreign technology and, upon having assimilated and improved on it, spin-off private firms to exploit the technology. Second, a working division of labour ensues, whereby ITRI/ERSO keep identifying and absorbing new technologies, infusing the private sector with them and providing the necessary assistance through their own R&D projects. A conceptual micro/meso link in the transition process allowing no less for truncation than for successful emergence seems missing.

This entailed conceiving and implementing a long-term capability-building strategy, including the formation of a critical mass of expertise to foster an indigenous industry. Globalisation during early to mid-1980s eased the way to a global outsourcing semi-conductor production market by ITRI/ERSO's main spinoffs (UCM and TSMC), which became key in the emergence of a full-fledged Taiwanese (supplier) electronics cluster. Underlying this was a clear new II-targeted policy component emanating from the highest level of the state, an example of strategic innovation policy.

Isenberg (2010).

For an illustration concerning the Korean machinery and equipment industries, see Lim (2012) in Lin (2012), 201–04.

Frequent changes in developing country macroeconomic and political regimes, including major alterations in policy priorities provoke further delays and disruptions. This is aptly illustrated by Argentina's failed attempts at catching up in aircraft, where efforts were mired in twists and turns owing to political vicissitudes (see Sercovich and Teubal 2012).

Lin (2012) puts forward an algorithm for choosing good II candidates for promotion. Although he does consider the comparative international catch-up experience, evolutionary traits and the role of the state, his proposal seems unduly simplistic. A more circumspect view of the matter put forward by Pack (2012, 196–201) relies on a highly stringent threshold relating to unlikely prior knowledge on spillovers, dynamic scale economies and coordination needs. Possessing such information in advance would undoubtedly be helpful but, unfortunately, not feasible for non-omniscient policymakers. Still, imperfect approximations can be tried to guide public policy.

Cumulative processes and indirect effects (e.g. failure in one instance may be a condition for success in a subsequent one) add to the intricacies involved. Malerba and Nelson (2012) examine various factors accounting for success and failure, such as sector-specific (technical and economic) characteristics and a number of other intervening actors and factors including vertical links, the knowledge system and policies. In this article we attempt to complete this approach by distinguishing learning at various levels of aggregation (the ‘learning triad’), analising such learning at, and accounting for, various critical points in the process dynamics involved, considering shifts to higher level systems/organisations, making explicit the role of policy learning and fully factoring-in fundamental uncertainty.

For the distinction between policy entrepreneurship and public entrepreneurship, see Groenewegen and van der Steen (2007) and Ostrom (2005).

See Sercovich and Teubal (2012), 358–59. On ‘cost discovery’, see Hausmann and Rodrik (2003).

Only a few top-tier industrialising economies have traversed successfully the pre-emergence phase in a critical mass of activities (e.g. South Korea, Taiwan, Singapore, Israel) enough to propel them to a ‘developed’ status. Others, like Brazil and India, have overcome it only in a few sectors (see Malerba and Nelson 2012). Most countries have stalled in this phase. For more details, see Sercovich and Teubal (2012), 362–63.

See Sercovich and Teubal (2012), 359.

For an illustration relating to the Brazilian iron and steel industry, see UNIDO (2005), 54.

See Avnimelech, Rosiello, and Teubal (2010).

Favourable exogenous variables may also help (e.g. the significant boost that deregulation of air traffic in the USA during the Reagan's administration meant to Embraer exports or the Oslo Peace process and the massive immigration from the former USSR in Israel during the early-1990s) and/or a new strategic programme (such as the original massive manpower programme in aeronautical engineering carried out by the Brazilian government, which enabled Embraer to start its successful trajectory, or Israel government's programme ‘Yozma’, which targeted a domestic venture-capital industry and market). For more details, see Sercovich and Teubal (2012), 359 and 364.

This system trait may not lead far enough in the absence progress by the other elements of the learning triad.

For example, by training post-graduates abroad in topics relating to the II promotion strategy envisaged, even though decisions on whether to actually materialise such intensions may still be pending.

While this paper focuses on full-fledged II policy targeting, there are other explicit policies as well as general economic measures that may indirectly facilitate/trigger/sustain the emergence of a new II either by design or otherwise.