The renaissance of industrial policy: Known knowns, known unknowns, and unknown unknowns

Governments around the world have rediscovered industrial policies as a means to accelerate development outcomes. Despite this growing interest, we still know surprisingly little about what works. In this blog, we try to summarize what we know, where we need more evidence, and how to design policies when their overall impact may remain uncertain.  
 

The renaissance of industrial policy

Between 2017 and 2023, the use of industrial policy increased nine-fold (figure 1). Advanced economies have driven most of this growth. In 2023 alone, two-thirds of new industrial policies were put in place by advanced economies. 

^{Sources: Global Trade Alert; Juhász et al. (2022, 2023); New Industrial Policy Observatory (NIPO, database); World Bank.}

But emerging market and developing economies (EMDEs) were not far behind: Almost all have increased the use of industrial policies. In 2023 alone, EMDEs implemented more than 500 new industrial policies.

Historical evidence suggests that industrial policies can tackle market failures, such as environmental pollution, and drive meaningful structural change—provided policies are implemented in the right enabling environments, as argued in the World Bank’s latest South Asia Development Update.

Yet, history has not been kind to industrial policy. The consensus that emerged from the industrial policies of the 1970s and 1980s is that, more often than not, these policies have failed.

Partly inspired by past lessons from East Asia and Latin America, contemporary industrial policies tend to be smaller in scale, as failed industrial policies in the past often had heavy fiscal costs.

In addition, industrial policies today are more outward-oriented and focus on integrating into global value chains. They are less likely to be driven by import substitution industrialization than they were in the 1970s to 1990s. Specifically, the use of tariffs has declined substantially. Import tariffs account for less than 1.2 percent of industrial policy interventions today. And average tariff rates have plunged since 1970 (figure 2). 

^{Sources: Lardy (2003); Rajaraman (2007); US International Trade Commission 2022; World Development Indicators (database); World Tariff Profiles (World Trade Organization 2023); World Bank.}

^{Note: Latest data is for 2021. Chart shows the effectively applied tariff rate. For China, data for 1970 is unavailable since the economy was fully state-controlled. Tariffs were only introduced in 1980. Here, data for China show tariff of 1982 from Lardy (2003).}

Similarly, the use of measures restricting global investment flows has declined. As illustrated by the widely used Chinn-Ito index, financial openness has increased dramatically since 1970, although far less in emerging markets than in advanced economies (figure 3).

The known knowns: Lessons from past industrial policy

Based on the lessons from past industrial policy, a few guiding principles emerge that can help increase the likelihood of success of industrial policy:

1. A focus on technology adoption
2. A focus on competition
3. An avoidance of trade-restricting measures
4. Transparency and monitoring.

Technology adoption. Industrial policies are more likely to succeed if they focus on the adoption of existing advanced technologies rather than innovation. Brazil, for example, taxed foreign intellectual property and subsidized domestic patenting in the early 2000s, as part of an innovation-driven growth strategy. The result was a host of low-quality patents and anemic growth.

South Asia, like other EMDEs, could increase the adoption of advanced technology from abroad. Indian firms, for example, have been enthusiastic adopters of basic technologies. About 80 percent of Indian firms have adopted energy efficient LED lighting, a much larger share than in other emerging markets. Where they have fallen short has been in advanced energy-efficient technologies, like for example programmable thermostats.

Competition. If targeted industries face market competition, the goals of industrial policy are more likely to be achieved. Greater openness to global trade and investment could help generate the productivity gains associated with competition while also expand access to larger markets.

The success of industrial policy in South Korea relied on intense competition in both domestic and international markets. In contrast, according to the World Bank’s latest Be-Ready Report, South Asian countries score in the bottom two quartiles by market competition among their peers.

Avoidance of trade-restricting measures. The need for competitive markets is also why industrial policy should avoid measures that restrict trade.

One alternative to trade-restricting measures is investment in infrastructure. While this may not be industrial policy in the traditional sense, it can boost specific sectors over others depending on the location and design of investment. In India, for example, the 3.4 percent of GDP infrastructure spending envisaged in the 2024-25 budget specifically targets renewable energy and digital infrastructure to boost the country’s global competitiveness.

Transparency and accountability. Industrial policies can be expensive, either because subsidies incur fiscal cost or because trade restrictions generate costly misallocation. Hence, industrial policy needs to be accompanied by transparent implementation and monitoring to curb favoritism and ensure accountability.
 

The known unknowns: Data gaps

The revival of policy interest in industrial policy is fairly recent and data are still lagging. Thus far, there is no publicly available database that would allow a comparison over time and across countries.

The New Industry Policy Observatory (NIPO) classifies and tracks industrial policies implemented or announced since 2023, but does not include policies implemented before 2023. Réka Juhász, Nathan Lane, and co-authors compare industrial policies during 2009-2020 based on a different classification method, but their database is not yet public.

The closest and most comprehensive proxy that has been published, courtesy of the Global Trade Alert, is a detailed database of all the trade measures that could potentially be considered industrial policy.

This comprehensive database gives a sense of the composition of EMDEs’ industrial policies. Consider, for example, India’s industrial policy measures implemented since 2011.

First, India’s policies have been more likely to be trade-related measures, including import and export tariffs and non-tariff measures. Localization measures, such as local content requirements, have been used similarly as in other EMDEs (figure 4). 

Second, corporate subsidies provided in India have been less transparent than elsewhere. Less than 5 percent of these subsidies were direct transfers, compared with about one-third in other emerging markets (figure 5). More than 80 percent of India’s subsidies were in the form of foregone revenues, which are much more difficult to measure and report than direct transfers. 

To transparently monitor the implementation, and assess the impacts of, the new generation of industrial policy, better data are much needed.
 

The unknown unknowns: The curse of partial equilibrium

Even with better data, it may be difficult ex ante to anticipate all the potential spillover effects of industrial policies. And because the policy by design chooses specific sectors or places to promote, there is inherent endogeneity. That is why empirical evaluation of industrial policy is particularly challenging, as Nathan Lane of Monash University argues in a recent literature review.

Contemporary work has confronted the challenge of endogeneity, by considering localized random variation in place-based policies, such as export promotion policies in Denmark and Chile, or exploiting discontinuities in policy design, such as the impact of R&D tax incentives on innovation in the U.K. and the U.S. These within-country studies also find mixed impacts of government interventions, but again they focus only on a small segment of the economy and cannot measure the aggregate impact of industrial policy.
 

Conclusion

It is therefore prudent to remain focused in the application of industrial policy and ready to adjust as more evidence emerges. Generating such evidence should be a high priority. And in the meantime, there is plenty of work for governments to make markets more open and competitive and invest in human capital and infrastructure, and thereby increase the chances of industrial policy success.

Reference

Chinn, M. D., and H. Ito. 2008.“A New Measure of Financial Openness.” Journal of Comparative Policy Analysis: Research and Practice 10 (3): 309–22.

Juhász, R., N. Lane, E. Oehlsen, and V. C. Pérez. 2022.“The Who, What, When, and How of Industrial Policy: A Text-Based Approach.” SSRN Scholarly Paper 4198209, Social Science Research Network, Rochester, NY.

Juhász, R., N. Lane, E. Oehlsen, and V. C. Pérez. 2023. “Updated Descriptive Statistics for ‘The Who, What, When and How of Industrial Policy: A Text-Based Approach.’” Working paper.

Lardy, N. R. 2003. “Trade Liberalization and Its Role in Chinese Economic Growth.” Working paper.

Rajaraman, I. 2007. “Trade and Investment: Tax Aspects.” Presentation delivered at the United Nations.