1 Economic development, economic complexity and industrial policy

Introduction

Sustained differences in economic growth between countries have led to vast differences in income per capita. When Adam Smith wrote The Wealth of Nations in 1776, the ratio of the highest to lowest income per capita was around 7 to 1. Today, this ratio is more than 250 to 1!⁽¹⁾See Bolt and Van Zanden (2020). Bolt, J. and J.L. Van Zanden (2020). Maddison style estimates of the evolution of the world economy: A new 2020 update. Maddison-Project Working Paper No. WP-15. University of Groningen. The historical roots of these significant disparities in income date back to the Industrial Revolution, when per capita income growth accelerated in industrialized countries.⁽²⁾For a more in-depth historical growth perspective, see Chapter 1 of WIPO (2015). World Intellectual Property Report 2015: Breakthrough Innovation and Economic Growth. Geneva: WIPO. Available at: www.wipo.int/publications/en/details.jsp?id=3995. Since then, continued technological progress has enabled new industries to grow. This growth has transformed the face of economies everywhere, though not in the same way.

While the gap between the poorest and richest countries has grown overall, the growth experience of initially poor economies has been mixed. Up until around 1990, poorer countries, taken together, did not grow any faster than richer countries. Economists characterized this development performance as a continued process of divergence.⁽³⁾As famously noted by Pritchett, there was “divergence, big time.” See Pritchett, L. (1997). Divergence, big time. Journal of Economic Perspectives, 11(3), 3–17. DOI: https//doi.org/10.1257/jep.11.3.3. This trend has, however, flipped over the other way during the last three decades, with poorer countries seeing somewhat faster growth than richer ones.⁽⁴⁾The seminal study pointing this out is Patel, D., J. Sandefur and A. Subramanian (2021). The new era of unconditional convergence. Journal of Development Economics, 152, 1026687. DOI: https//doi.org/10.1016/j.jdeveco.2021.102687. In other words, divergence has become convergence.

Despite this better news, it is important to realize that what holds true for poorer economies as a whole does not necessarily hold true for all such economies. Already by the 1970s and 1980s, certain less developed countries – notably, the Republic of Korea – had achieved economic convergence with developed economies. Conversely, despite there being a general convergence trend over the last three decades, a considerable number of poorer economies – including many of the least developed countries – have struggled to generate growth and have continued to fall behind. Across world regions, Asian economies – primarily those in East and South Asia – have, overall, seen a stronger growth performance than economies in Africa and Latin America. Yet even within different regions, the growth experience of the different economies has varied over time.

Economists have sought to explain such wide variations in development outcomes for as long as they have been observed. Robert Solow famously argued that long-run economic growth could only be achieved through sustained technological development.⁽⁵⁾Solow, R.M. (1956). A contribution to the theory of economic growth. Quarterly Journal of Economics, 70(1), 65–94. DOI: https//doi.org/10.2307/1884513. Joseph Schumpeter as well as Philippe Aghion and Peter Howitt emphasized the importance of creative destruction, whereby new technologies and industries replace old ones.⁽⁶⁾Schumpeter, J.A. (1942). Capitalism, Socialism and Democracy. Harper and Brothers; Aghion, P. and P. Howitt (1992). A model of growth through creative destruction. Econometrica, 60(2), 323–351. DOI: https://doi.org/10.2307/2951599. Paul Romer devised a theory of so-called endogenous growth, in which technological progress attracts investments in both human capital and research and development (R&D).⁽⁷⁾Romer, P.M. (1990). Endogenous technological change. Journal of Political Economy, 98(5, part 2), S71–S102. Martin Weitzman argued that new ideas arise through the recombining of existing ones and that economic growth is constrained not by a lack of new ideas, but the inability to leverage existing ones.⁽⁸⁾Weitzmann, M.L. (1998). Recombinant growth. The Quarterly Journal of Economics, 113(2), 331–360. DOI: https//doi.org/10.1162/003355398555595.

In principle, technological knowledge can be easily shared and replicated, but in practice it does not flow seamlessly across space.⁽⁹⁾Maskell, P. and A. Malmberg (1999). The competitiveness of firms and regions: ‘Ubiquitification’ and the importance of localized learning. European urban and regional studies, 6(1), 9–25. DOI: https//doi.org/10.1177/096977649900600102; Cohen, W.M. and D.A. Levinthal (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 128–152. DOI: https//doi.org/10.2307/2393553; WIPO (2015). World Intellectual Property Report 2015: Breakthrough Innovation and Economic Growth. Geneva: WIPO. Available at: www.wipo.int/publications/en/details.jsp?id=3995. Some places have historically been more successful at acquiring knowledge and effectively converting it for industrial use than have others.

What explains their relative success and what can others learn from their experiences?

These questions are at the heart of this report. Drawing on new thinking in economic research, the report introduces the concept of economic complexity, which provides a framework for understanding how economies accumulate, diversify and apply knowledge. This task is performed in Section 2 of this chapter. Based on this understanding, Section 3 reviews what policymakers can do to promote industrial growth and reviews how both the practice and the intellectual thought behind so-called industrial policy have evolved over the past few decades. Section 4 offers concluding thoughts by highlighting the emergence of a new era of industrial policies.

The remainder of this report delves deeper into the industrial development process. Mirroring the notion of economic complexity, Chapter 2 employs data on trade, scientific publications and patents to develop indicators of so-called innovation complexity. Chapters 3 to 5, in turn, offer insights into successful industrial development approaches through three case studies: agricultural technology, motorcycles and videogames.

Understanding relatedness and economic complexity

Technology represents the knowledge that we harness to reshape our physical and social environments. It has grown tremendously over the past few centuries, as demonstrated by the ever increasing volume of books, scientific papers and patents. Yet our individual capacity to comprehend it remains constrained. Hence, we increasingly become specialized as individuals and distribute knowledge to counterparts. Over time, this knowledge ends up in tools, machines and equipment – so-called embodied knowledge. In addition, we codify what we know and convert it into forms that can be shared through documentation, standardization and classification – so-called codified knowledge. Yet a large part of our knowledge is harder to codify; instead, it remains tacit.

Tacit knowledge is “stuck” in brains and does not easily move across the world. Even codified knowledge does not flow seamlessly from one individual to another, because it requires prior knowledge to absorb. As a result, knowledge becomes concentrated in certain places. In 2021, for instance, Germany exported 61.5 percent of the world’s stereoscopic microscopes and the United States of America (US) 60.4 percent of the world’s aircraft launching gear (Figure 1.1). Firms and workers in those industries are highly specialized and cannot easily switch from producing microscopes to aircraft gear or vice versa.

This concentration and specialization of tacit knowledge can also be observed within industries. For instance, companies that make jet engines typically do not produce other aircraft parts. To put all the parts of an aircraft together, somebody must think about the design of each part and how they will come together. Hence, the growth of knowledge at the product level requires an increase in the division of labor at the level of the individual.

The game of Scrabble as a metaphor

In this metaphor,⁽¹⁰⁾See Haussmann, R., M.A. Yildirim, C. Chacua, M. Hartog and S. Gadgin Matha (2024). Innovation policies under economic complexity. WIPO Economic Research Working Paper No. 77. World Intellectual Property Organization. goods and services rely on productive capabilities, which we can call letters. Words then represent the combination of those productive capabilities that go into making a particular product. Not all combinations of letters are words: some sets of letters are words, but other sets of letters are just gibberish. So the products – words – can be defined by the set of capabilities needed to make them – the letters.

Continuing the metaphor, we can think of places as collections of words and letters and products as collections of letters. A place – whether a city, province or  country – can be characterized by the letters it has, whereas a product can be characterized by the letters it requires.

Economic researchers have formalized these logical statements about the world and have tested them empirically.⁽¹¹⁾Hidalgo, C.A., B. Klinger, A.-L. Barabási and R. Hausmann (2007). The product space conditions the development of nations. Science, 317(5837), 482–487. DOI: https//doi.org/10.1126/science.1144581; Haussmann, R., C.A. Hidalgo, S. Bustos, M. Coscia, A. Simoes and M.A. Yildirim (2014). The Atlas of Economic Complexity: Mapping Paths to Prosperity. The MIT Press. Using international trade data, they found that the difference in the number of letters explains not only which products a given place is likely to diversify into but also the pattern of diversification. This holds true for countries and municipalities alike. They visualize this in what they call the product space (Figure 1.2). Each dot in this space represents a product. The proximity between the dots approximates the similarity in the know-how required to produce two products. Lines connecting two dots indicate the primary connections between products. The product space is irregular, with different products bunched together. For example, garment products are tightly clustered together, implying that the letters needed to make one kind of garment are similar to the letters needed to make other kinds of garments. The same goes for machinery. Then, some poorly connected products suggest that those words are short. For instance, oil requires making holes in the ground, but there are few products for which one needs to dig holes in the ground. By comparison, letters to make a microwave oven are similar to the letters to make a washer or dryer.

The world's product spaceFigure 1.2 Exported products clustered and connected based on common related capabilitiesNotes: Each dot represents a product category (based on the Harmonized System 1992 classification) and the size of each dot is proportional to the size of that product's world trade. The colors represent the 10 major sectors shown in the key (textiles, agriculture, stone, minerals, metals, chemicals, vehicles, machinery, electronics, and others). Products requiring related capabilities are clustered closer together in the network. The lines between dots indicate primary connections between products.Source: Harvard Growth Lab (2023); an interactive version is available at https://atlas.cid.harvard.edu.

Each economy occupies its own product space. In order to diversify into new products, a country needs to find ways to reach other parts of the product space. And research has shown that diversification favors activities that are close together – known as the principle of relatedness.⁽¹²⁾Neffke, F., M. Henning and R. Boschma. (2011). How do regions diversify over time? Industry relatedness and the development of new growth paths in regions. Economic Geography, 87(3), 237–265. DOI: https://doi.org/10.1111/j.1944-8287.2011.01121.x; Boschma, R. (2017). Relatedness as driver of regional diversification: A research agenda. Regional Studies, 51(3), 351–364. DOI: https://doi.org/10.1080/00343404.2016.1254767; Hidalgo, C.A., P.-A. Balland, R. Boschma, M. Delgado, M. Feldman, K. Frenken, E. Glaeser, C. He, D.F. Kogler, A. Morrison, F. Neffke, D. Rigby, S. Stern, S. Zheng and S. Zhu (2018). The principle of relatedness. In Morales, A., C. Gershenson, D. Braha, A. Minai and Y. Bar-Yam (eds), Unifying Themes in Complex Systems IX: Proceedings of the Ninth International Conference on Complex Systems. Cham: Springer, 451–457. DOI: https://doi.org/10.1007/978-3-319-96661-8_46; Hausmann, R., D. Stock and M.A. Yildirim (2022). Implied comparative advantage. Research Policy, 51(8), 104143. DOI: https//doi.org/10.1016/j.respol.2020.104143; Li, Y. and F. Neffke (2023). Evaluating the principle of relatedness: Estimation, drivers and implications for policy. CID Research Fellows and Graduate Student Working Paper Series No. 146. Cambridge, MA: Center for International Development at Harvard University.Neffke and Henning apply this principle to job transitions, with people moving to jobs that have similar skill requirements. See Neffke, F. and M. Henning (2013). Skill relatedness and firm diversification. Strategic Management Journal, 34(3), 297–316. DOI: https//doi.org/10.1002/smj.2014.  For example, Venezuela has relatively few activities and most of them – such as oil and raw materials – are in peripheral positions.⁽¹³⁾See Figure 6 in Haussmann, R., M.A. Yildirim, C. Chacua, M. Hartog and S. Gadgin Matha (2024). Innovation policies under economic complexity. WIPO Economic Research Working Paper No. 79. World Intellectual Property Organization. Mexico has more, and Austria – which is less than a 10th the size of Mexico – has even more, covering the entire product spectrum.

Further implications of the Scrabble metaphor

The Scrabble metaphor has further implications for industrial development. First, the more letters one has, the more words one can put together. The diversity of letters thus leads to a diversity of words. Second, the longer the word, the harder it is to make it. If one assumes that there is some distribution of letters among places, the longer the word, the fewer the places that can make it. We call the number of places that can make a product the ubiquity of the product. A place that has a lot of letters should be more diversified and specialized in less ubiquitous products. And by the same token, less ubiquitous products should be made in more diversified places. Accordingly, the term economic complexity seeks to capture the knowledge residing in an economy, as it is expressed in the diversity and ubiquity of the products it produces. Economic researchers have translated this concept into a formal economic complexity index that captures how diversified and ubiquitous an economy’s export basket is.⁽¹⁴⁾Hidalgo, C.A. and R. Hausmann (2009). The building blocks of economic complexity. Proceedings of the National Academy of Sciences, 106(26), 10570–10575; Haussmann, R., C.A. Hidalgo, S. Bustos, M. Coscia, A. Simoes and M.A. Yildirim (2014). The Atlas of Economic Complexity: Mapping Paths to Prosperity. The MIT Press.

The economic complexity index correlates with economic output and income per capita. Poor countries have few letters, whereas rich countries have a lot of letters. More importantly, if – relative to other countries with the same GDP or income per capita – a country has fewer letters, it tends to grow more slowly. In some sense, the space of letters is more fundamental to growth prospects than the current income level. A greater endowment with letters enables economies to diversify more easily and especially into products that are less ubiquitous. This is discussed in more detail in Chapter 2.

The concept of economic complexity allows for a richer understanding of the economic growth process that emphasizes different dimensions of knowledge and how distributive knowledge needs to come together in order to make things. To sustain growth, countries need to move to new and more complex industries, while building on relevant local capabilities.

A key question therefore is how can a country acquire new letters to make new words and diversify into new activities, particularly toward the denser part of the product space, where more complex products are located. The country would need the knowledge to make such jumps. According to the economic complexity view, the key to economic development is productive knowledge. Such knowledge is distributed in different individuals’ heads, tools and materials, and the process of economic growth entails the accumulation and expression of this knowledge in more goods. In the Scrabble world, this corresponds to more letters, more words and longer words. Development thus requires greater specialization at the level of individuals, which leads to greater diversification at the level of companies and industries.

One implication of this theory is that the tacit and complex knowledge embedded in specialized individuals is crucial for diversification. Due to its tacit and complex nature, such knowledge does not move freely across the world. Hence, countries may lack the knowledge to make jumps into new activities. How can they possibly overcome this barrier?

Diversification is a chicken-and-egg problem. For a place to diversify into new activities, it must learn to do things that it could not do in the past. But how does a place begin to make things if it does not know how they are made? For example, how does one become an experienced watchmaker in a place that does not make watches? The greater the number of missing letters, the more challenging it will be to diversify.

Countries generally need to benefit from a high level of basic skills to be able to absorb more complex and specialized capabilities from elsewhere.⁽¹⁵⁾Hanushek, E. and L. Woessman. (2015). The Knowledge Capital of Nations: Education and the Economics of Growth. MIT Press.  Still, the accumulation of basic skills also faces the chicken-and-egg problem. In a location where most of the industries present do not use high-skilled or specialized labor, the incentives to invest in these skills are limited. Only once there are concrete prospects from acquiring new skills will these incentives change. In the complexity framework, we can think of letters as being shaped by the education system; if a letter has no word in which to be used, it will not emerge.

The complexity framework developed here is in line with the criticism by economists – notably Benjamin F. Jones – about how human capital stock is accounted for in studies of economic growth.⁽¹⁶⁾Jones, B.F. (2014). The knowledge trap: Human capital and development reconsidered. Northwestern University Working Paper. Available at: www.kellogg.northwestern.edu/faculty/jones-ben/htm/Knowledge%20Trap.pdf.  In particular, Jones highlights the importance of skills that are not perfect substitutes for one another. A heart surgeon, for instance, requires at least an anesthesiologist to operate effectively – without one, the surgeon’s value would greatly decrease. The division of labor allows for the existence of collective know-how, which is greater than the sum of individual skills. In developed economies, the ubiquity of other highly specialized knowledge workers justifies the costly acquisition of specialized knowledge. Developing economies, in turn, may find themselves in a “knowledge trap,” as an insufficient ecosystem for complementary skills undermines an individual’s reward for investing in specialized skills.⁽¹⁷⁾Jones empirically accounts for the specialized nature of and complementarities between skills and finds that an adjusted measure of human capital stock fully explains the large income variations between rich and poor economies. See Jones, B.F. (2014). The human capital stock: A generalized approach. American Economic Review, 104(11), 3752–3777. DOI: https://doi.org/10.1257/aer.104.11.3752.

How then can the challenge of missing letters be overcome? One way is the presence of major organizations with diversified portfolios, which allows for internal diffusion and redeployment of capabilities. Particularly relevant here are those organizations with substantial resources – such as corporate conglomerates – that can re-deploy existing workers and even whole teams to new and related activities. Famous examples of such internal diversification are the keiretsu in Japan and the chaebols in the Republic of Korea. Research has shown how these organizations propelled diversification into new technological activities in these two economies.⁽¹⁸⁾Saxenian, A. (2006). The New Argonauts: Regional Advantage in a Global Economy. Harvard University Press.

Labor mobility leading to knowledge transfer

Worker migration is another way of enabling diversification.⁽¹⁹⁾See, for example, Miguelez, E. and A. Morrison (2022). Migrant inventors as agents of technological change. The Journal of Technology Transfer, 48, 669–692. DOI: https://dx.doi.org/10.2139/ssrn.4256518. Take the case of East Germany, which has experienced a gradual revival of its industries after initially losing 60 percent of its manufacturing jobs following German unification. Research has shown that the pioneer plants in East Germany relied heavily on experienced workers from West Germany.⁽²⁰⁾Hausmann, R. and F.M. Neffke (2019). The workforce of pioneer plants: The role of worker mobility in the diffusion of industries. Research Policy, 48(3), 628–648. DOI: https//doi.org/10.1016/j.respol.2018.10.017. These well-paid workers from outside the region generated substantial employment opportunities for local workers and individuals entering the job markets. Pioneer plants, in turn, may train workers which may then be hired by follower plants, ultimately fostering the diffusion of specialized knowledge.

Pioneer plants also tend to drive the structural transformation of economies. Research has found that the greatest diversification steps are taken by entrepreneurs and existing firms from elsewhere that set up new plants in regions. They foster a process of knowledge diffusion across regions. During a period of 17 years of structural transformation in regions within Sweden, it is those pioneers that are shown most likely to survive and thrive.⁽²¹⁾Neffke, F., M. Hartog, R. Boschma and M. Henning (2018). Agents of structure change: The role of firms and entrepreneurs in regional diversification. Economic Geography, 94(1), 23–48. DOI: https://doi.org/10.1080/00130095.2017.1391691. By contrast, existing firms that try to jump into new activities are more likely to fail. They cannot draw on related local capabilities in the same way that pioneer companies can draw on internal capabilities acquired elsewhere.

Hence, for countries and regions to achieve structural transformation, it may be important to attract companies and workers from elsewhere. Economist AnnaLee Saxenian calls these workers the New Argonauts. She documents how foreign-born, highly skilled workers who have ventured back and forth between Silicon Valley and their home countries infused the latter with new knowledge. ⁽²²⁾Saxenian, A. (2006). The New Argonauts: Regional Advantage in a Global Economy. Harvard University Press. Saxenian finds that such interactions proved crucial to emerging innovation hotbeds, such as the semiconductor industry in Taiwan, Province of China. More generally, these interactions are reflected in patterns of world trade. One study finds that a 10 percent increase in immigration from exporters of a given product leads to a two percent increase in the probability that the host country starts exporting the same product within the next 10 years.⁽²³⁾Bahar, D. and H. Rapoport (2018). Migration, knowledge diffusion and the comparative advantage of nations. The Economic Journal, 128(612), F273–F305. DOI: https://doi.org/10.1111/ecoj.12450.

One driving force for labor mobility is foreign direct investment (FDI). When companies invest abroad, they often send experienced workers to their subsidiaries and transfer skills to local employees through formal training and mentoring. FDI can thus be a key steppingstone to acquiring new letters. One study, for example, found that when companies set up foreign R&D facilities in a particular region, that region sees subsequent growth in patenting activities.⁽²⁴⁾Crescenzi, R., A. Dyèvre and F. Neffke (2022). Innovation catalysts: how multinationals reshape the global geography of innovation. Economic Geography, 98(3), 199–227. DOI: https://doi.org/10.1080/00130095.2022.2026766. More generally, research has shown that most structural change is induced not by domestic firms but by foreign ones.⁽²⁵⁾Elekes, Z., R. Boschma and B. Lengyel (2019). Foreign-owned firms as agents of structural change in regions. Regional Studies, 53(11), 1603–1613. DOI: https://doi.org/10.1080/00343404.2019.1596254. That said, how the knowledge of multinational companies diffuses through the host economy depends critically on the pre-existing capabilities of local firms.⁽²⁶⁾See Crescenzi, R., A. Dyèvre and F. Neffke (2022). Innovation catalysts: how multinationals reshape the global geography of innovation. Economic Geography, 98(3), 199–227. DOI: https://doi.org/10.1080/00130095.2022.2026766. The benefits of FDI thus vary across industries and countries.

Diaspora networks can be another important channel of knowledge diffusion. They enable countries to tap into the knowledge of natives living abroad. For example, studies have traced the origins and growth of the information technology industries in China, India and Israel back to professional connections between domestic engineers, on the one hand, and diaspora engineers and entrepreneurs in Silicon Valley, on the other.⁽²⁷⁾Smart, A. and J.-Y. Hsu (2004). The Chinese diaspora, foreign investment and economic development in China. The Review of International Affairs, 3(4), 544–566. DOI: https://doi.org/10.1080/1475355042000241511; Saxenian, A. (2006). The New Argonauts: Regional Advantage in a Global Economy. Harvard University Press; Pandey, A., A. Aggarwal, R. Devane and Y. Kuznetsov (2006). The Indian diaspora: A unique case? In Kuznetsov, Y. (ed.), Diaspora Networks and the International Migration of Skills. World Bank Institute Development Studies, 71–97. Similarly, the recent modernization of agriculture in Albania and its growth in exports of agricultural products can be traced back to returning migrants from Greece and Italy, who brought in advanced technological know-how.⁽²⁸⁾Hausmann, R. and L. Nedelkoska (2018). Welcome home in a crisis: Effects of return migration on the non-migrants’ wages and employment. European Economic Review, 101, 101–132. DOI: https://doi.org/10.1016/j.euroecorev.2017.10.003.

These various mechanisms focus on moving brains rather than moving knowledge across brains. As discussed above, the tacitness of knowledge makes the latter much harder to accomplish. One way it can still work is through business travel. Despite modern communication technology, in-person travel remains an important feature of global business activity. One study found that business travel networks predict which new economic activities will develop in a country and, inversely, which old activities will decline.⁽²⁹⁾Coscia, M., F.M. Neffke and R. Hausmann (2020). Knowledge diffusion in the network of international business travel. Nature Human Behaviour, 4(10), 1011–1020. DOI: https://doi.org/10.1038/s41562-020-0922-x. In particular, business travel from countries specializing in a specific industry causes growth in that economic activity in the destination country. In fact, the study finds that this effect, in statistical terms, has the most substantial impact on a range of bilateral relationships between countries, such as FDI, trade and migration.

The acquisition and accumulation of specialized knowledge usually result from market forces, with individuals and businesses identifying opportunities to maximize wages and profits. However, the diverse outcomes in industrial development observed across different parts of the world and over time indicate that the learning process enabling industrial diversification is not automatic. This raises the question of what preconditions need to be in place for successful learning to occur and, in particular, which public policies favor such success? We turn to this question next.

Policies to promote industrial development

From a broad perspective, a wide range of preconditions and policies matter for industrial development. For instance, overall macroeconomic stability, a functioning legal system, an effective educational system, and a financial system that efficiently turns savings into investment are all important. The complexity framework just discussed also highlights the importance of labor mobility, particularly openness to skilled immigration.

No doubt it is possible to point at countries that did not fully meet all the above conditions and yet still experienced industrial diversification. In addition, successful industrial development can be a self-reinforcing process, whereby initial success and the resultant economic growth promote the preconditions for subsequent success. Still, unfavorable overall preconditions in the above-listed areas will, on balance, hinder industrial development prospects.

Looking at it more narrowly, governments have long implemented policies aimed at directly promoting industrial diversification. The array of policy instruments employed for this purpose has significantly evolved over the past half-century. Traditionally, so-called industrial policy has been associated with a range of policy measures – notably, import tariffs, subsidies and subsidized loans – targeted at a limited set of industries.

An explicit or implicit aspect of these industrial policies was the selection of winners – the belief that certain industries held more promise for future development prospects compared to others. Starting in the 1950s, industrial policy centered on import substitution strategies, with many developing economies in Asia, Africa and Latin America pursuing such strategies.⁽³⁰⁾For a comprehensive review see Irwin, D.A. (2021). The rise and fall of import substitution. World Development, 139, 105306. DOI: https://doi.org/10.1016/j.worlddev.2020.105306. The logic behind import substitution was that infant industries in less developed economies needed temporary protection that would enable them to learn and achieve scale, before becoming globally competitive.

The track record of import substitution is mixed, and in the 1990s many economists turned against it. While a theoretical case for import substitution exists, there were increasing doubts whether governments can reliably predict which industries hold the greatest promise. In addition, the acquisition of productive knowledge often proved more formidable than anticipated and perpetual infancy led to enduring import protection.⁽³¹⁾Bruton, H.J. (1998). A reconsideration of import substitution. Journal of Economic Literature, 36(2), 903–936.

The disenchantment with import substitution gave rise to a new industry policy paradigm: export-led growth. The impetus behind this strategy was the success observed in certain East Asian economies, most notably Japan, the Republic of Korea, Hong Kong, SAR and Singapore, collectively recognized as the East Asian miracle. These economies experienced simultaneous industrial diversification and a significant expansion of their export activities. This led to the idea that active engagement in global commerce played a pivotal role in fostering knowledge acquisition and promoting industrial diversification. In principle, policies aimed at promoting export-led growth were intended to be neutral toward specific industries, avoiding the picking of winners. The emphasis was on flexible exchange rates, diminishing import protection, and, broadly, eliminating distortions in market incentives.⁽³²⁾Striving for export-led growth was part of the so-called Washington Consensus, as summarized by Williamson, J. (1990). What Washington means by policy reform. In Williamson, J. (ed.), Latin American Adjustment: How Much Has Happened? Washington, DC: Institute for International Economics, 7–20. However, in reality, some government policies, such as offering special incentives within export processing zones, still favored certain sectors at the expense of others.

Many governments still subscribe to the export-led growth paradigm. However, economists' perspectives on its likelihood of success have significantly evolved. The outcomes of initial export-led growth policies exhibit a mixed track record. Despite the notable East Asian success story, many developing economies that adopted these policies did not experience substantial industrial growth.⁽³³⁾Fink, C. and J. Raffo (2019). What role for intellectual property in industrial development? In Correa, C. and X. Seuba (eds), Intellectual Property and Development: Understanding the Interfaces. Springer Singapore, 119–136. DOI: https://doi.org/10.1007/978-981-13-2856-5. At the very least, the track record suggests that these policies are insufficient to generate industrial diversification.

Economic research has also raised doubts about whether the East Asian model of export-led growth, which is focused on manufacturing, can be effectively replicated in today's vastly transformed global economy.⁽³⁴⁾See Juhász, R., N.J. Lane and D. Rodrik (2023). The new economics of industrial policy. National Bureau of Economic Research Working Paper No. 3153. Available at: www.nber.org/papers/w31538. The manufacturing sector now contributes significantly less to both economic output and employment compared to the period when the East Asian miracle occurred. Modern industrial development strategies must encompass the growth of service sectors, as these sectors typically account for most economic output. Although technology has expanded the tradability of some services, a substantial portion of service activities remains non-tradable, with limited opportunities for learning through exporting.

A related question is whether all countries and regions share the same potential for industrial diversification. For example, advanced economies already specializing in highly complex activities may be in a better position to diversify into other highly complex activities, whereas less developed economies will only be able to diversify into less complex activities.⁽³⁵⁾For evidence of such inequality-reinforcing feedback loops, see Pinheiro, F.L., P.-A. Balland, R. Boschma and D. Hartmann (2022). The dark side of the geography of innovation: Relatedness, complexity and regional inequality in Europe. Regional Studies, Special Issue. DOI: https://doi.org/10.1080/00343404.2022.2106362. The potential for diversification may also change over time, as technological progress opens new opportunities for diversification and closes others.

Considering the insufficiency of the export-led growth paradigm, how has thinking on industrial policy evolved over the past two decades? And does the concept of economic complexity help in setting priorities for industrial policy? While a full review of the rich research on industrial policy is beyond the scope of this chapter, the remainder of this section focuses on two prevalent lines of thought that have defined more recent approaches to industrial policy: harnessing science, technology and innovation for industrial development, and effectively devising industrial policy.

Harnessing science, technology and innovation for industrial development

A country's science, technology and innovation (STI) system comprises all entities engaged in the creation and dissemination of knowledge, as well as the interactions between them. These entities include universities, training institutes, research organizations, regulatory institutions and companies, which can be publicly or privately owned and can operate on either a for-profit or non-profit basis.

Within the economic complexity framework, an STI system shapes the set of letters available in a place and, in turn, the number of words – or products – it can produce. While productive knowledge ultimately resides in the brains of skilled workers, these workers are often trained at local universities. In addition, when innovating, companies frequently collaborate with scientific organizations, drawing in expertise not available in-house. Scientific organizations – while seeking the advancement of scientific knowledge – are often expressly tasked to address the technological needs of the local economy, though some do so more successfully than others.⁽³⁶⁾Balland and Boschma show how the match between scientific capabilities and technological needs varies across European regions. See Balland, P.-A. and R. Boschma (2022). Do scientific capabilities in specific domains matter for technological diversification in European regions? Research Policy, 51(10), 104594. DOI: https://doi.org/10.1016/j.respol.2022.104594.

Public policies and funding play a crucial role in sustaining an STI system that supports industrial development. Markets left to themselves would systemically underinvest in generating and diffusing knowledge. This is most evident for scientific research that does not have any immediate industrial application, but which provides the foundation for future innovations. Yet even where there are prospects for industrial applications, private markets may shun them. Technology may still be incipient, with a high risk of failure and uncertain commercial viability. There are numerous examples of technological advances initially benefiting from public funding that spawned entirely new industries, such as the pharmaceutical innovations developed during the Second World War, the internet and self-driving cars.⁽³⁷⁾For historical perspectives on the origins of various breakthrough innovations, see Mazzucato, M. (2015). The Entrepreneurial State: Debunking Public vs. Private Sector Myths (Revised edition). New York, NY: PublicAffairs; WIPO (2015). World Intellectual Property Report 2015: Breakthrough Innovation and Economic Growth. Geneva: WIPO. Available at: www.wipo.int/publications/en/details.jsp?id=3995; and WIPO (2022a). World Intellectual Property Report 2022: The Direction of Innovation. Geneva: WIPO. Available at: www.wipo.int/wipr/en/2022.

Even when companies are willing to bear the risk of investing in innovation, economists contend that they often tend to underinvest compared to what is socially desirable, for two key reasons. First, companies may find it difficult to appropriate the returns on their innovations if others can readily copy them. This is why governments protect patents and other forms of intellectual property (IP), which offer companies timebound exclusive rights on their inventive and creative outputs. Second, even when companies are able to profit from their innovations, the private return on innovation is often substantially below its social return.⁽³⁸⁾Jones, B.F. and L.H. Summers (2020). A calculation of the social returns to innovation. National Bureau of Economic Research Working Paper No. 27863. DOI: https://doi.org/10.3386/W27863. Take the example of COVID-19 vaccine innovation. One study compared the social return from the invention of these vaccines – in the form of saved lives and contained economic output losses – to the private profits accruing to vaccine makers. It estimated the former to exceed the latter by a factor of 887.⁽³⁹⁾Fink, C. (2022). Calculating private and social returns to COVID-19 vaccine innovation. WIPO Economic Research Working Paper No. 68. World Intellectual Property Organization. Available at: www.wipo.int/publications/en/details.jsp?id=4595&plang=EN. Higher social than private returns justify governments providing extra incentives for companies to invest in innovation. Such incentives take the form of R&D subsidies and tax credits, subsidized loans, prizes and other instruments.⁽⁴⁰⁾For an overview of these instruments, see Table 2.2 in WIPO (2022a). World Intellectual Property Report 2022: The Direction of Innovation. Geneva: WIPO. Available at: www.wipo.int/wipr/en/2022.

An STI system is not only important in fostering innovation that prompts the commercialization of technology that is new to the world. It can also be instrumental in enabling economies to absorb and adapt knowledge generated elsewhere. In fact, this will be the primary role of STI systems in developing economies, where industries do not operate at the world’s technology frontier. For example, in many developing economies, universities and research institutes have led the adaptation of new plant varieties and farming technologies to local conditions. Studies have also documented the crucial role that publicly-funded research organizations played at the early stages of India’s pharmaceutical industry and the Republic of Korea’s semiconductor industry.⁽⁴¹⁾See Mazumdar, M. (2013). An overview of the Indian pharmaceutical sector. In Performance of Pharmaceutical Companies in India. Berlin: Springer, 17–44; and Kim, S.R. (1998). The Korean system of innovation and the semiconductor industry: A governance perspective. Industrial and Corporate Change, 7(2), 275–309. DOI: https://doi.org/10.1093/icc/7.2.275, respectively. The support provided by these organizations included advice on technology deployment, the transfer of technology developed within such organizations, joint R&D, and other services.

Harnessing STI systems for development is a matter of mobilizing government revenue to fund universities and public research organizations and providing R&D support to companies. However, it is also a matter of providing policy incentives, building linkages within the STI system and offering advice. For instance, many universities have developed frameworks for transferring technology developed in academic labs to companies. These frameworks seek to promote such transfers, while also recognizing that the commercial deployment of university technology often requires substantial follow-on investment by companies. Managing IP rights is a critical component of such technology transfer frameworks.

In addition, STI institutions are important in matching the supply and demand for technology. Unlike many goods and services traded in the marketplace, technology is highly differentiated. Companies may not be aware of existing solutions to the technological challenges they face, while academic researchers may be insufficiently informed about the technological needs of companies. There is thus a role for platforms and industry fairs that can overcome such informational divides. Similarly, many IP offices provide so-called technology landscapes to industry stakeholders, based on the patents filed in different technology fields worldwide (Box 1.1).

Finally, through the STI system, governments can prioritize the development and diffusion of technologies that societies value highly but for which private incentives to innovate are insufficient. One example is public health. The COVID-19 pandemic has highlighted the importance of an agile innovation system to respond to the spread of infectious diseases. However, in non-pandemic times, companies have little incentive to invest in vaccine R&D, as there is no market to sell vaccines. Public support can compensate for the lack of market incentives. Similarly, STI policies can prioritize technologies that reduce carbon emissions and better manage climate adaptation – beyond what market incentives can provide.

Devising industrial policy

As pointed out above, one criticism leveled at early industrial policies was the uncertain ability of governments to predict which industrial activities held the greatest promise in a particular location. There are, indeed, numerous examples of disappointing results from initiatives to build, say, local biotech or semiconductor industries from the ground up.⁽⁴⁵⁾See Lerner, J. (2009). Boulevard of Broken Dreams: Why Public Efforts to Boost Entrepreneurship and Venture Capital Have Failed – And What to Do About It. Princeton University Press; and, as a specific example, Orsenigo, L. (2001). The (failed) development of a biotechnology cluster: The case of Lombardy. Small Business Economics, 17(1/2), 77–92. Policymakers arguably overestimated the capacity of local economies to build the knowledge base necessary to enable the production of globally competitive goods, especially in industries with rapid technological progress and fast product cycles. Indeed, the economic complexity framework emphasizes the key role of the existing capabilities available within a territory, which limit the possibilities of developing new technologies.⁽⁴⁶⁾Hidalgo, C.A., P.-A. Balland, R. Boschma, M. Delgado, M. Feldman, K. Frenken, E. Glaeser, C. He, D.F. Kogler, A. Morrison, F. Neffke, D. Rigby, S. Stern, S. Zheng and S. Zhu (2018). The principle of relatedness. In Morales, A., C. Gershenson, D. Braha, A. Minai and Y. Bar-Yam (eds), Unifying Themes in Complex Systems IX: Proceedings of the Ninth International Conference on Complex Systems. Cham: Springer, 451–457. DOI: https://doi.org/10.1007/978-3-319-96661-8_46. Knowledge diversification is typically a gradual path-dependent process, where one decision determines later ones. In the Scrabble metaphor, it is driven by exploiting adjacent opportunities to combine letters to form new words.

Industrial policy interventions may serve different purposes. They may amplify existing capabilities and thus accelerate industrial diversification. Alternatively, they may aim to disrupt the natural path dependence process if existing capabilities constrain industrial growth.⁽⁴⁷⁾Balland, P.-A., R. Boschma, J. Crespo and D.L. Rigby (2019). Smart specialization policy in the European Union: Relatedness, knowledge complexity and regional diversification. Regional Studies, 53(9), 1252–1268. DOI: https://doi.org/10.1080/00343404.2018.1437900; Pinheiro, F.L., P.-A. Balland, R. Boschma and D. Hartmann (2022). The dark side of the geography of innovation: Relatedness, complexity and regional inequality in Europe. Regional Studies, Special Issue. DOI: https://doi.org/10.1080/00343404.2022.2106362. Either way, the formulation of industrial policies needs to rely on a careful understanding of existing skills and capabilities, their competitive strengths and opportunities for diversification and growth. New approaches to industrial policy have thus emphasized the need for industrial policies to be devised in a “bottom-up” rather than a “top-down” fashion, typically taking place at the level of specific economic regions rather than overall economies.

One approach – adopted in the European Union (EU) since 2014 – is called smart specialization.⁽⁴⁸⁾See Foray, D. (2015). Smart Specialisation: Opportunities and Challenges for Regional Innovation Policy. Abingdon: Routledge; and the European Union’s Smart Specialization Platform at: https://s3platform.jrc.ec.europa.eu. This entails an inclusive process that involves stakeholders and is centered on “entrepreneurial discovery.” The process seeks to identify priorities for investment by both governments and companies that build on local capabilities. For example, governments invest in physical infrastructure, research and educational institutes that address the specific needs of local entrepreneurs. Companies, in turn, invest in innovation that continuously sustains their competitive strength. The key principles of smart specialization policies include focusing on the most binding constraints to industrial growth and instituting constant evaluation and calibration. Smart specialization policies also acknowledge the important role of STI systems in fostering the acquisition and diffusion of knowledge that enables industrial diversification.

How successful has smart specialization been? It is difficult – and perhaps too early – to provide a definite answer. The first phase of implementation in EU countries suggests that it is feasible but challenging to design and implement smart specialization strategies. In many cases, it proved difficult to identify a relevant priority area. Moreover, translating identified priorities into specific goals and finally moving toward transformational roadmaps and activities requires supporting institutions, which often proved to be weak.

Assessing economic outcomes is complicated by the fact that smart specialization strategies entail numerous interventions over time with long-term objectives, which do not allow for straightforward “before” and “after” assessments.⁽⁴⁹⁾Foray, D. (2019). In response to ‘Six critical questions about smart specialization’. European Planning Studies, 27:10, 2066–2078. DOI: https://doi.org/10.1080/09654313.2019.1664037. While recent evidence suggests that EU cities with the largest gains in complex and related technologies have enjoyed an economic performance premium, it remains an open question to what extent smart specialization policies can actually foster such gains.⁽⁵⁰⁾Rigby, D.L., C. Roelser, D. Kogler, R. Boschma and P.-A. Balland (2022). Do EU regions benefit from Smart Specialisation principles? Regional Studies, 56(12), 2058–2073. DOI: https://doi.org/10.1080/00343404.2022.2032628.

A new era of industrial policies

In summary, economists have long held ambivalent views about the effectiveness of industrial policies. This arguably reflects the mixed historical record of industrial policies centered on import protection and subsidizing specific industries. Beneath this ambivalence, however, there arguably is a more widespread consensus on the types of policy intervention that promote industrial development. Most economists would endorse governments investing in an STI system that facilitates the acquisition and diffusion of new knowledge.⁽⁵¹⁾In reviewing industrial policy carried out in the United States from 1970 to 2020, Hufbauer and Jung draw a similar conclusion. Hufbauer, G.C. and E. Jung (2021). Scoring 50 years of US industrial policy, 1970–2020. Peterson Institute for International Economics. Available at: www.piie.com/publications/piie-briefings/scoring-50-years-us-industrial-policy-1970-2020.

STI policies may not always fit neatly within narrow definitions of industrial policy, and they typically fall outside the purview of industrial development ministries. However, STI systems invariably shape industrial development and are usually designed with this purpose in mind. To be clear, STI policy choices still entail taking tough decisions. For example, which fields of scientific research should receive support? Which incipient technologies deserve funding and when should such funding be withdrawn? Market mechanisms provide few, if any, signals to inform such decisions.

Recent years have seen a revival of industrial policies, including in economies that have traditionally shunned them, such as the United States. This revival was not prompted by economists having gained new insights into their effectiveness. Rather, they emerged as a response to various new challenges faced by governments. One such challenge is climate change and the need to reduce carbon emissions. Prominently, the European Green Deal of 2020 and the U.S. Inflation Reduction Act of 2022 provide incentives to companies and projects that promote the development and deployment of carbon-reducing technology. Another challenge is to address shortages of strategic goods in the face of global supply chain shocks – as happened during the COVID-19 pandemic. For example, the Government of Japan has made subsidies available to companies to support the re-shoring of production from overseas locations.⁽⁵²⁾See www.jetro.go.jp/en/invest/attractive_sectors/manufacturing/government_initiatives.html. Finally, some governments have rolled out large-scale support to certain high-tech industries considered critical for national security, notably semiconductors.⁽⁵³⁾See, for example, the U.S. Chips and Sciences Act of 2022.

The industrial policy instruments employed by governments in this new era vary widely. They range from tax breaks, production subsidies and R&D subsidies to trade and regulatory measures.⁽⁵⁴⁾Recent industrial policies implemented in a wide range of jurisdictions are documented in Evenett, S., A. Jakubik, F. Mart��n and M. Ruta. (2024). The return of industrial policy in data. IMF Working Paper No. 2024/001. International Monetary Fund. Available at: www.imf.org/en/Publications/WP/Issues/2023/12/23/The-Return-of-Industrial-Policy-in-Data-542828. Most policy measures are still being implemented, so it is too early to assess their overall impact. Moreover, while recent industrial policies may aim to foster industrial development as a complementary goal, their success must be evaluated in the context of the broader objectives these policies intend to achieve.

Nonetheless, and despite the difference in context, some of the choices in recent industrial policies reflect bets on which industrial activities will offer long-term benefits. For example, many governments around the world currently support the development of an indigenous battery industry, believing that batteries will be a critical input for electrical vehicles and, more generally, future mobility. However, economies of scale in the production of batteries may mean that production will only be efficient at a few locations. In addition, it is unclear whether hosting battery production will offer substantial spillover benefits to the local economy, or whether batteries will turn out to be a commoditized product that can be easily imported. Notwithstanding the strong imperative behind the new wave of industrial policies, evaluating the benefits and costs associated with different policy interventions will thus remain important.