Industrial Policy Design – A tale of ‘Giant Gnomes’ and ‘Illusory Giants’!

Research under WP4 was mainly undertaken by Imperial College London, LSE, and University of Maastricht. It considered what approach governments should pursue when designing a strategy for Industrial and R&D policy. The outcome is a suggestion that governments should build their strategy around clearly identified and measured market failures. This research has especially considered one particular key market failure, knowledge spillovers, which imply that innovating firms do not invest enough in R&D because the social returns are higher than the private ones.

Hence, governments can increase productivity by specifically targeting particular sectors, technology areas or firms generating higher social returns. To operationalise this, the researchers developed, through the WATSON project, a new indicator based on global patent and stock market data called the Industrial Strategy Index (IStra-X)

IStra-X takes account of both direct and indirect spillovers. This allows the identification of Giant Gnomes, i.e. innovations whose immediate impact is dwarfed by their indirect impact, as well as Illusory Giants, innovations whose immediate impact is small compared to their indirect impact. It is shown that some of the most impactful innovations are Giant Gnomes.

A methodology was devised to estimate private returns to innovation. Private return estimates are used as a key input to estimate the external value of innovations and, by extension, the Industrial Strategy Index (IStra-X) reported in the WATSON report D4.3 Social Returns Report. The methodology takes into account variations in value of innovations between different technology fields. IStra-X was computed for various key countries at the level of detailed technology areas. This provided a suggestive ranking of areas for governments to focus on.

To infer the private return of an innovation, the researchers relied on the extrapolation of estimates to the relevant population of inventions. To do this, they used information on technological classification, time, geographic location and scope of protection that is available for (nearly) all patented innovations. Results suggest that the extrapolation models—whilst leaving room for improvement—capture a considerable amount of heterogeneity in the private returns to innovation. This methodology is outlined in the WATSON Report D4.1 Returns to Innovation Report.

The WATSON report D4.3 Social Returns Report describes the researchers’ approach to addressing therenewed appetite for vertical industrial policy and strategy in advanced industrialized countries. The motivation is typically a desire to promote economic growth. The researchers under WP4 suggest that a possible mechanism to guide strategy is the large and persistent differences in the knowledge spillovers generated by different technologies.

By directing innovation support to technologies and sectors generating the highest spillovers, governments can have a positive effect on growth. The researchers developed a new way to measure innovation spillovers from patent and balance sheet data called Patent Rank (Prank). This builds on Google’s Page Rank but instead of a probability measure it provides a value measure of spillovers. It takes into account both direct and indirect spillovers and differentiates between global and national spillovers. It is argued that the latter are relevant for a national industrial strategy.

It is also shown that there is significant and economically meaningful variation in the value of spillovers generated by different technology areas. The framework also uncovers considerable variation between different countries, between measures of comparative advantage and measures of spillovers as well as between national and global spillovers.

There is also considerable variation in the amount of internalisation of spillovers at the country level; e.g. the US, Germany, and Japan are internalising spillovers that are nearly twice as valuable as the spillovers internalised by countries such as the UK or France. Taken together, the results leave little hope for a one-size-fits-all industrial strategy, but the proposed framework allows for a tailored approach to better address market failures implied by positive externalities from knowledge spillovers.

Link: D4.1 Returns to Innovation Report

Link: D4.3 Social Returns Report