Title: The Who, What, When, and How of Industrial Policy: A Text-Based Approach 

 Abstract: Whether industrial policy (IP) can promote economic change is an age-old question in economic development and growth. Yet, outside a handful of case studies, we understand little about these policies—including basic facts. Two key reasons for this deficit are lack of measurement and lack of data. Our paper addresses these issues by creating a text-based approach to measuring industrial policy and providing a global view of practice in the 21st century. We use machine learning to construct a classifier for coding instances of IP from textual summaries of policy. We then apply our tool to a comprehensive, English-language database of international economic policy. In doing so, we create a new dataset of industrial policy at the country-product-year level, and establish five fundamental facts about global industrial policy from 2009 to 2020. First, IP is practiced by a very limited set of countries. Second, the countries engaged most in IP tend to be wealthier (in the top income quintile) liberal democracies, and IP is rare among the poorest nations (bottom quintile). Third, when countries use IP, they do so selectively: policies tend to be highly targeted towards a small set of industries. Rich and poor countries target similar sectors, with the exception of green IP (wealthier countries) and textiles (poorer countries). Fourth, industrial policy has indeed returned, and practice has trended upward since the 2010s. Fifth, countries use very specific tools for IP aims. Many trade policies are not industrial policy measures, and among the many potential instruments, the bulk of IP are subsidies and export-promotion measures. Thus, we find contemporary practice is a far cry from industrial policy’s past, and tends toward specific, selective, export-oriented policies used by the world’s more developed economies.

Marta Santamaria will be hosting this visit.