Guest post: Factory fit
If industrial policy is back, how can we make it more effective?
After decades of neglect, industrial policy is back in vogue. ‘Industrial policy is so hot right now’, the Financial Times recently declared, while The Economist called it a ‘revival’ and called for some history lessons based on past misadventures. Harvard economist, Ricardo Hausmann, linked the renewed enthusiasm for industrial policy to the end of the Reagan-Thatcher revolution and the Washington Consensus, while other popular explanations focus on perceived post-pandemic trends, such as deglobalization and ‘reshoring’, the Green Revolution and escalating geopolitical rivalries.
If this sense of a historical revival is correct, it begs the question: how can this period of industrial policy improve on past precedents? While debates rage on relentlessly over the relative importance of industrial policy in the post-war growth miracles of, for example, the Asian Tigers (notably Taiwan and South Korea), examples of failure very likely exceed the success stories (see here and here). In short, government-directed efforts at shaping the structure of the economy by identifying promising sectors and industries and then cultivating them through combinations of public investment, tax incentives, subsidies and trade protections, have a decidedly mixed track record.
In two recent papers, published through Stanford and MIT, respectively, my co-authors and I argue that the design, implementation and evaluation of industrial policy faces a set of practical challenges. Which market failures are depriving otherwise promising growth sectors of private capital? How and why would industrial policy overcome these market failures? When should sector-specific supportive policies be updated or redirected, given that the success of such policies ultimately lies in their removal from competitive and self-sufficient industries? How do we assess the effectiveness of industrial policies, ex ante and ex post?
Our research proposes that the analytical framework and annually updated dataset on ‘Economic Fitness’, developed and maintained by the World Bank, can assist in exactly these tasks (we focus on applications and use cases by Sovereign Development Funds, as a particular instrument of industrial policy, but the implications are applicable to industry policy writ large).
Like the by now well-established literature on Economic Complexity, Economic Fitness captures an economy’s level of diversification and its ability to produce complex products through observed patterns in export data. The assumption, which enjoys strong empirical support based on the work of Ricardo Hausmann and Dani Rodrik, is that the structure of exports proxies the existence of productive capabilities. The ‘fittest’ economies are not only specialized in complex productive activities, but are also diversified across activities: they are able to produce the largest bundles of products, ranging from the simplest to the most complex, using the largest set of productive capabilities. Productive capabilities are not independent of each other, but rather situated in a complex network structure (the differences between Economic Fitness and Economic Complexity are overwhelmingly methodological, rather than conceptual or theoretical).
What does this mean in practice? A corollary of the network structure connecting ‘nodes’ of productive capabilities to each other is that the cultivation of one set of capabilities catalyses the emergence of adjacent or related ones. Once the evolution and cultivation of productive capabilities are understood in this manner, the potential for market failures and correction through industrial policy becomes evident: if private investors do not fully capture the broader social benefits from expanding the network of productive capabilities, the aggregate level private investment in capability-enhancing activities will be suboptimal.
Private capital may flow towards areas of economic activity that may be profitable in the short run, but not towards those that are socially or strategically optimal in the sense that they improve the overall ‘fitness’ of an economy. Time horizons matter too: it might take a decade or possibly even longer – well beyond the typical investment horizon of private investors – to successfully cultivate the types of self-reliant economic activities (and their associated productive capabilities) that meaningfully diversify an economy and catalyse the gradual emergence of ever-more complex related products, services and sectors.
The World Bank’s Economic Fitness database yields an information-rich toolkit that has been successfully integrated with the Bank’s investment evaluations in recent years. While our papers discuss these data and their potential uses in real-world policy and institutional settings in detail, a brief overview of the toolkit and its applications provide a flavour of what we have in mind:
The ‘Product Progression Network’ maps the network-based pathways to higher Economic Fitness, as experienced by other countries. The network structure reflects the ability of the framework to identify, with a high degree of statistical significance, which products are connected through a causal relationship (success in one product therefore providing signals to policymakers as to likely related areas of economic activity).
‘Progression Probability’ or ‘Feasibility’ score represents the probability of cultivating competitiveness in a specific sector over a five-year horizon, given existing productive capabilities, which are identified through competitiveness in exports, mapped onto the network structure.
‘Fitness Gain’ scores are a probabilistic expression of the expectation that success in one sector will unlock opportunities to upgrade to related products/sectors/industries in the future. This can help policymakers weigh up the expected benefit – expressed as the contribution to overall or sector-specific ‘fitness’ (recall, this is the ability to produce both a range of goods and services competitively, as well as having competitive and capabilities in the most complex activities).
These and other Economic Fitness tools can provide an empirically driven and theoretically sound basis for targeting and periodically updating industrial policies. Mapping sectors and industries according to Feasibility and Fitness Gains, for example, can help policymakers target an appropriate combination of public and private capital, with private capital being incentivised towards high-feasibility areas, while public capital is reserved for areas that may have somewhat lower feasibility but promise high fitness gains. Economic Fitness can help identify productive-capability shortages, which can subsequently be addressed through education policy, efforts to attract Foreign Direct Investment and strategic cross-border investments. Finally, network mapping and feasibility indicators can guide policymakers in periodically (at least annually) analysing whether targeted industrial support can be removed and redirected. If competitiveness and self-reliance have been established in one industry or sector, policy targets can shift to related ones that have been newly enabled (measured through a rise in feasibility).
Does Economic Fitness work in all settings across all time periods? Of course not. For economies with limited productive capability endowments, it can be hard to find many compelling areas of feasibility to target (empirically, Economic Fitness performs better out of sample for middle-income countries than low-income ones). Moreover, factors exogenous factors, such as terms-of-trade shocks, commodity prices and geopolitics, can dominate Economic Fitness indicators in the short- to medium-term. That said, we are hopeful that consistent applications of the tools and framework of Economic Fitness can lead to improvements over the decidedly mixed historical track record of industrial policy.
Photo by Karsten Winegeart on Unsplash.