Introduction and Overview

Abstract

The world economy was severely hit by the COVID-19 pandemic in 2020. It was estimated that the annual growth in the world’s real gross domestic product (GDP) in 2020 would be $-3.3$% (International Monetary Fund (IMF) 2021). World trade simultaneously contracted sharply.

The world economy was severely hit by the COVID-19 pandemic in 2020. It was estimated that the annual growth in the world’s real gross domestic product (GDP) in 2020 would be $-3.3$% (International Monetary Fund (IMF) 2021). World trade simultaneously contracted sharply. It was estimated that the growth in the world’s trade volume of goods and services would be $-8.5$% in 2020 (IMF 2021). According to the statistics released by the Johns Hopkins Coronavirus Resource Center, by the end of 2020 the cumulative number of infected people worldwide is over 83 million and the cumulative number of worldwide deaths is over 1.81 million. However, the decline in global manufacturing was short-lived and both advanced and emerging economies showed V-shaped recoveries in manufacturing output in the second half of 2020 (IMF 2021, Fig. 1.1.1). Moreover, thanks to vaccines and various policy supports, the world economy is projected to grow at 6% in 2021 (IMF 2021), but still faces great uncertainty. Its recovery depends on the path of the health crisis.

Although the COVID-19 pandemic proved that the world economy is vulnerable to health risks, at the same time it shows its good adaptability. The size of the COVID-19 recession is expected to be smaller than the 2008 Global Financial Crisis (IMF 2021). Its influences are different from one country to another. Low income countries with limited capacity for policy support were hit relatively harder than advanced economies. Furthermore, countries that rely on tourism and commodity exports were particularly severely damaged. Those countries are expected to suffer more significant medium-term losses. Using the pandemic vulnerability index (PVI), which is calculated by national data on COVID-19 morbidity and mortality rates and other related information, Shrestha et al. (2020) show that certain countries are more vulnerable to the COVID-19 pandemic than others. According to their analysis, the top 10 highly vulnerable countries include Brazil, India, the United States, Russia, South Africa, Chile, Mexico, Iran, Peru, and Pakistan.

It is argued that globalization was a major driving force behind the fast spread of COVID-19 from China to the rest of the world. For example, Farzanegan et al. (2021) show that countries with higher levels of socio-economic globalization are exposed to a higher case fatality rate due to COVID-19, according to the KOF Globalization Index (the ratio of confirmed deaths to confirmed cases), covering more than 150 countries in July 2020. Shrestha et al. (2020) argue that “trade and travel, essential components of globalization, are significant contributors to the spread of infectious diseases” (p. 1). Historically, pandemics have repeatedly emerged together with human activities and movements.¹ By reviewing the history of pandemic influenza, Saunders-Hastings and Krewski (2016) argue that pandemic influenza is a consequence of human development and that globalization in relation to human behavior, demographics, and mobility has enhanced the threat of pandemic emergence and accelerated the spread of novel viruses. Conversely, they point out that globalization has also facilitated international cooperation in disease prevention, control, and treatment by promoting advances in disease research and surveillance.

In this book we pay attention to the proliferation of regionalism from the mid-1990s and the globalized activities of multinational enterprises (MNEs) as important elements of recent globalization and explore interactions between these two elements of globalization and diffusion of knowledge in the world. Diffusion of knowledge across countries is important because it affects the speed at which the world’s technology frontier expands. For example, Eaton and Kortum (1996) show that more than 50% of the economic growth in 19 advanced countries in the 1980s derived from innovation in the United States, Japan, and Germany. Moreover, diffusion of knowledge contributes to income convergence across countries (Keller 2004).

This chapter starts by presenting the background of the study in this book and then provides an overview of the book.

1.1 Background of the Study

1.1.1 Globalization and the Proliferation of Regionalism

The term “globalization” is commonly used, but it means different things to different people. Globalization in the economic sense means the “integration of national economies into international economy” (Bhagwati 2004, p. 3) through international trade, foreign direct investment (FDI), and international flows of workers and technology. Alternatively, economic globalization can be defined as “the increased interdependence of national economies, and the trend towards greater integration of goods and factor markets” (Neary 2003, p. 246). There has been a heated debate about the pros and cons of globalization among economists (e.g., Bhagwati 2004; Rodrik 1997, 2011, 2018; Samuelson 2004; Stiglitz 2002, 2006, 2018). For example, Samuelson (2004) uses a Ricardian model and numerical examples to illustrate the possibility of a country suffering a welfare loss from a trading partner’s productivity growth in the country’s export good sector. But he supports globalization by arguing that “free trade may turn out pragmatically to be still best for each region in comparison with lobbyist-induced tariffs and quotas which involve both perversion of democracy and nonsubtle dead-weight distortion losses” (p. 143). Whereas Bhagwati (2004) defends globalization by answering many criticisms from the anti-globalism side over the issues of its impacts on poverty, child labor, culture, labor standards, and the environment, Stiglitz (2002, 2018) emphasizes that globalization has been mismanaged and argues what should be done to make globalization more equitable.

The current wave of globalization is not the first. The world economy reached a peak of globalization just before World War I, when trade and FDI attained for then unprecedented levels (Deardorff and Stern 2002). Baldwin (2006, 2011, 2016a) explains the waves of globalization by the theory of “unbundling.” According to him, the first unbundling, which is the unbundling of production and consumption across national borders, occurred when the transportation revolution—railroads and steamships—dramatically lowered transport costs in the first half of the nineteenth century. Since then, until around 1990, countries engaged mainly in trade in final consumption goods, according to their comparative advantage as traditional international trade theory such as the Ricardian model and Heckscher-Ohlin model predicted, and experienced gains from international trade.² Baldwin calls the first unbundling “old globalization.” Then, the second unbundling, which is the “spatial unbundling of production stages previously clustered in factories and offices” (Baldwin 2011, p. 5), was derived from the information and communication technology (ICT) revolution around 1990. Not only transport costs but also communication costs were substantially reduced, so that stages of production that previously had to be performed in close proximity could be performed in geographically distant locations. Due to the second unbundling, production processes were fragmented, FDI in production facilities increased, and trade in parts and intermediate goods was greatly expanded. A theory of fragmentation developed by Jones and Kierzkowski (1990) explains these changes in production and trade.³ Baldwin calls the second unbundling “new globalization.” Moreover, the third unbundling, which is the unbundling of tasks to individuals located in different countries due to a reduction in face-to-face communication (Baldwin 2016a, 2019), may have already started. Further advances in both information technology (IT) and communication technology (CT) will lead to this third unbundling. “Telemigration” (i.e., virtual presence of foreign workers through the advancement of CT) and “globotics” (i.e., a combination of globalization and a new form of robotics by the advancement of IT such as artificial intelligence) characterize the third unbundling (Baldwin 2019). Baldwin (2016a) calls the third unbundling “future globalization.”

Fig. 1.1

Source: METI (2019), Fig. II–1–1–1–9

Bilateral trade accounting for over 0.1% of the value of global trade (2000). Notes: This figure is created from the IMF’s Direction of Trade Statistics. Trade between Hong Kong and other countries is excluded.

Fig. 1.2

Source: METI (2019), Fig. II–1–1–1–10

Bilateral trade accounting for over 0.1% of the value of global trade (2017). Notes: This figure is created from the IMF’s Direction of Trade Statistics. Trade between Hong Kong and other countries is excluded.

The trend of globalization after 2000 can be seen by comparing bilateral trade in the world between 2000 and a recent year (2017). Figures 1.1 and 1.2 show the amount of bilateral trade in 2000 and 2017, respectively (Ministry of Economy, Trade, and Industry (METI) 2019). Comparing these two figures, the changes in the hub countries of bilateral trade within eighteen years are significant. In the figures, countries in a blue and red circle are a developed and an emerging/developing country, respectively. A red-filled circle represents bilateral trade that accounts for over 0.1% of global trade, and it exceeds $1 trillion in total. A blue-filled circle represents bilateral trade, and it exceeds $500 billion in total, and a green-filled circle represents bilateral trade, and it exceeds $100 billion. Blue lines indicate ties between developed countries, red lines represent ties between emerging/developing countries, and green lines ties between developed and emerging developing countries. The thickness of lines between two countries represents the size of the total trade amount on a scale, from the thickest to the thinnest, of (1) over $200 billion, (2) over $100 billion, (3) over $50 billion, and (4) below $50 billion.

From the figures we can observe the following five transitions (METI 2019):

• (1) The number of large-scale bilateral trades between emerging and developing countries around China expanded, and the trade in itself greatly increased in amount.

• (2) The network of bilateral trade in the East Asian region became dense, and the trade simultaneously increased in amount. Particularly, Vietnam was a rising country in trade.

• (3) The number of large-scale trade and the amount of trade within the European Union (EU) expanded.

• (4) The center of trade in the East Asian region moved from Japan to China.

• (5) Economic linkage between the East Asian region and the North American region strengthened.

We next turn to the issue of trade policy in the progress of globalization. Deardorff and Stern (2002) argue that both steady increases in international trade and international capital flows in the second half of the twentieth century, which are much of what has come to be called globalization, were caused by technology and policy. Baldwin’s theory of unbundling mainly focuses on changes in technology. On the other hand, policies that have enhanced both trade and investment are multilateral trade liberalization through the General Agreement on Tariffs and Trade (GATT)/the World Trade Organization (WTO) after World War II and the recent proliferation of regional trade agreements (RTAs) from the mid-1990s.⁴

Baldwin (2016b) illustrates how countries have succeeded in liberalizing trade through the GATT rounds of negotiations after World War II.⁵ Mainly, advanced countries, such as the United States, Western European countries, and Japan, reduced their import tariffs and non-tariff barriers until the start of the WTO in 1995. By contrast, the trade negotiation at the Doha Round that started in 2001 has been deadlocked. For the last two decades, little progress has been made on multilateral trade liberalization at the WTO. Baldwin (2016b) argues that the most commonly cited cause of the WTO’s difficulties is “the lost dominance of the advanced economies” (p. 106). The share of major advanced countries in world imports declined due to the rapid growth of emerging economies. At the same time, the sheer number of developing country members has shifted power in the WTO and made negotiations more difficult. In addition, Baldwin argues that regionalism and unilateral tariff-cutting by developing countries also created challenges to multilateral trade liberalization through the WTO. RTAs involve tariff cutting that would otherwise have had to be achieved through the WTO. Moreover, many of the new RTAs are “deep” in the sense that “they went beyond tariff-cutting and included legally binding assurances aimed at making signatories more business-friendly to trade and investment flows from other signatories” (Baldwin 2016b, p. 107). On the other hand, an expansion in offshoring from advanced economies opened a new pathway to industrialization through joining an international production network and expanding the amount and range of tasks performed (Baldwin 2016b). Since tariffs hinder rather than help industrialization in this new development model, developing countries started to cut their import tariffs unilaterally, independently of the WTO negotiations.

Because of the malfunctioning of the WTO in the 2000s for trade liberalization and rule setting, RTAs have been playing a more important role in the world economy than before.⁶ With regard to the impact of RTAs on bilateral trade, existing studies have obtained very different estimates. Cipollina and Salvatici (2010) investigate by a meta-analysis why the ex post measurements of the trade impact of RTAs are volatile. For this research, they use 1,827 point estimates of the impact of RTAs on bilateral trade from 85 studies (38 published journal articles and 47 working papers) and run meta-analysis regressions. After filtering out the publication impact and other biases, they find a robust, positive trade impact of RTAs equivalent to an increase in trade of around 40%. Since the estimates tend to become larger for more recent years, they argue that the tendency could be a consequence of the recent evolution from “shallow” to “deep” integration.

1.1.2 Global Firms and Production Networks in the East Asian Region

For the last two decades, empirical studies on international trade have provided evidence of firm heterogeneity in trade, as Melitz (2003) demonstrates theoretically (e.g., Bernard and Jensen 1995, 1999; Bernard et al. 2007, 2009, 2012, 2018). A large number of studies that employ micro data in many different countries have suggested that “global firms” play a dominant role in each market. Bernard et al. (2018) define “global firms” as “firms that participate in the international economy along multiple margins and account for substantial share of aggregate trade” (p. 566).

Bernard et al. (2018) use US firm and trade transactions data and show that only a subset of firms participate in international markets. These trading firms indicate superior performance characteristics: they are larger and more productive than other non-trading firms. Moreover, a large fraction of firms that export or import actually engage in both exporting and importing. “More successful firms export more of each product to each market, export more products to each market, export to more markets, import more of each product from each source country, import more products from each source country, and import from more source countries” (Bernard et al. 2018, p. 607). Global firms are likely to be MNEs. Bernard et al. (2009) report that US-based MNEs mediate more than 90% of US trade. Consistent with the model of heterogeneous firms (Helpman et al. 2004), Yeaple (2009) shows that more productive US firms tend to own affiliates in a larger number of countries and that these affiliates generate greater revenue from sales in their host economies.

Studies of Japanese firms have also provided evidence consistent with theories of heterogeneous firms (e.g., Head and Ries 2003; Kimura and Kiyota 2006; Todo 2011; Wakasugi 2014; Wakasugi and Tanaka 2010, 2012). Thus, global firms are likely to hold a dominant position in the globalized activities of Japanese firms. In some of the chapters in this book, we employ micro data on Japanese firms and focus on their globalized activities. Since Japanese MNEs play an important role in production networks in East Asia, it is worth looking at the situation of the supply networks in this region, as the background of the study.

According to METI (2019), about two thirds of annual intra-regional exports (of raw materials, intermediate goods, and final goods) in East Asia from 2011 to 2017 were shared by intermediate goods, whereas their export ratios in the 1990s were a little over one half. Annual intra-regional exports amounted to $1,400–1,600 billion in 2011–2017, while they ranged between $170 and $400 billion in the 1990s.⁷ Both amount and ratio of intra-regional exports have increased greatly for the 1990–2017 period. If we focus on the machinery industry in East Asia, in which the international division of labor is most developed,⁸ the ratio of intermediate goods in its intra-regional export during 1998–2017 was at 60–65% except for 2012, while its ratio in 1990 was less than 50% (METI 2019).⁹ In contrast, the amount of its trade increased rapidly from $40 billion in 1990 to about $830 billion in 2017. These results unambiguously substantiate that the production network is formed in the machinery industry in East Asia.

The system of typical international division of labor between the United States (developed country) and Mexico (developing country) focusing on intra-firm transactions was established by around the year 1990 (Ando and Kimura 2014). Later, this system developed into the one including East Asia in the machinery industry (mainly electrical and electronic sectors): parts and intermediated goods are exported from East Asia to Mexico, and then final goods and parts that are manufactured in Mexico are exported to the United States and Canada. After the Central and Eastern European countries—such as Poland, Czech Republic, Slovakia, Hungary, and Romania—became EU members in the fifth enlargement of the EU, industrial clusters were accelerated in these countries. As a result, the supply of machinery parts and intermediates from East Asia rapidly expanded (Ando and Kimura 2013). Among others, the import of electrical and electronic parts and intermediates increased remarkably. The production network between the EU and East Asia was consolidated via the Central and Eastern European countries (Ando and Kimura 2013). These countries have an active role as a catalyst, like Mexico in North America. The fact that East Asia is deeply related to the other two global production networks attracts our interest.

The enlargement of production networks in a region makes it possible to supply goods efficiently, while it may make the supply of goods vulnerable to shocks. In fact, it has been shown that production networks in East Asia are relatively resilient to severe shocks (Ando and Kimura 2012; Obashi 2011; Todo et al. 2015).¹⁰ Specifically, Ando and Kimura (2012) analyze the impact of the 2008–2009 Global Financial Crisis and of the 2011 Great East Japan Earthquake on Japanese exports, focusing on the characteristics of domestic/international production networks in machinery industries. They show that these two massive shocks generated common and different adjustments in production networks and trade. In the face of such severe shocks, trade within the production networks of machinery final products and other products demonstrates distinctive stability and resiliency. However, the magnitude and duration of the shocks were fairly different: the impact of the Financial Crisis was huge and prolonged, whereas that of the Japan Earthquake was much smaller and more temporary. They argue that the cause of such a difference is that the Financial Crisis was primarily a demand shock in the US and EU markets, while the Japan Earthquake was a supply shock due to the destruction of production plants in the impacted area.

The effects of the 2011 Great East Japan Earthquake on supply chain networks within Japan are examined by Todo et al. (2015).¹¹ Using firm-level data, they show that extensive supply chain networks are not always harmful to disaster recovery. More specifically, they find that having more suppliers and customers outside of the disaster area tends to shorten the recovery time, though it affects sales growth in the medium term only weakly. By contrast, having more suppliers and customers in the disaster area has no effect on the recovery time but tends to improve medium-term sales growth. In addition, they identify a negative effect from supply chains on recovery through the disruption of supply and demand and two positive effects through support and substitution. Overall, they conclude that the positive effects from extensive supply networks typically outweigh the negative effects, resulting in a net positive effect. Moreover, Obashi (2011) investigates the resilience of international production network in the Asian region to the 1997–1998 Asian Currency Crisis. She conducts a series of survival analyses and finds that transactions of machinery parts and components within the production network are more likely to be stable and resilient to a temporary disruption, compared to transactions of finished products. More specifically, during the Asian Currency Crisis, machinery parts and components were more likely to be traded through long-lived trade relationships than finished products. Besides, many of the trade relationships for machinery parts and components were restored shortly after the break caused by the Asian Currency Crisis, as compared to those for finished products.

1.1.3 Innovation and Diffusion of Knowledge

It is Joseph A. Schumpeter (1883–1950) who first asserted the importance of innovation for economic development in industrial society. These days, it is well recognized that innovation has a significant influence on the rise and fall of an enterprise. After his assertion, it has been extensively investigated and discussed in the literature whether firm size and market power have an effect on firm innovation. Schumpeter (1942) argued that a large-scale establishment is the most powerful engine of progress and that firms in concentrated markets have a stronger incentive to invest in innovation. Many theoretical and empirical studies have explored the relationship between market structure and innovation. Specifically, the Schumpeterian endogenous growth models, the pioneering work of which is Aghion and Howitt (1992), formalize Schumpeter’s argument. In contrast to the Schumpeterian theory, Arrow (1962) argues that a firm’s gains from innovation at the margin are larger in an industry that is more competitive ex ante. Assuming that property rights over invention are fully protected, he shows that a monopolist that is not exposed to competition under both old and new technologies has less incentive to invest in research and development (R&D) for a process innovation than does a firm in a competitive sector. Blundell et al. (1999) examine the relationship between a firm’s ex ante market power and innovation and find that the market share has a positive effect on innovation, whereas that of overall market concentration is negative, suggesting that, while a higher market share stimulates innovation, concentrated industries may innovate less. Many other empirical studies find a positive relationship between competition and innovation.¹² To reconcile the Schumpeterian theory with empirical evidence, Aghion et al. (2005) develop a simple model in which competition discourages laggard firms from innovating but encourages neck-and-neck firms in innovating and derive an inverted-U shaped relationship between competition and innovation. They provide strong empirical support for their theory using data on publicly listed manufacturing firms in the United Kingdom. In contrast, Hashmi (2013) finds evidence of a mildly negative relationship between competition and innovation from the US data. He modifies the model of Aghion et al. (2005) in such a way that the average technology gap is higher in the country where the relationship is negative and then show that the modified model can explain both negative and inverted-U shaped relationships. Some empirical studies provide evidence to support the Schumpeter’s hypothesis more strongly. For example, focusing on publicly traded US industrial firms in the 1910s and 1920s, Nicholas (2003) finds that both firm size and market power have significantly positive effects on patenting and, moreover, that financial markets reward firms for their innovative behavior with increasing their market values. He shows that all of these effects worked strongly during the 1920s.

Innovation is one of the important sources of competitive advantage for global firms (e.g., Aw et al. 2011; Atkeson and Burstein 2010; Costantini and Melitz 2008). Those firms make enormous investments in R&D. Then, both global firms and other firms benefit from diffusion of knowledge (Keller and Yeaple 2013; Sampson 2016). A large number of existing studies have shown that even for highly advanced economies like the United States, the outcomes of R&D in foreign countries play an important role in its own technical progress. Most of the other countries in the world are far more dependent on foreign R&D (Sveikauskas 2007). Therefore, innovation and diffusion of knowledge are both strongly related to globalization.

A common measure of the state of innovation is the number of patent applications. Patent applications worldwide were 3.224 million in 2019 (World Intellectual Property Organization (WIPO) 2020). A breakdown of the total applications by country is as follows: China was top-ranked, with 1.4 million applications and a 43.4% share of the world total, followed by the United States (0.621 million, 19.3%), and Japan (0.308 million, 9.6%). China’s share has increased considerably over the last decade from 17% in 2009 to 43.4% in 2019. Although Japan’s share has decreased from 18.8% in 2009 to 9.6% in 2019, Japan is still ranked in the top three.

In terms of the share of patent applications by region, Asia accounts for 65.0% (50.9%) in 2019 (2009), followed by North America 20.4% (26.6%), and Europe 11.3% (17.4%) (WIPO, 2020). Over the past 10 years, Asia’s share increased greatly by 14% points, whereas those of North America and Europe decreased by 6.2% points and 6% points, respectively. The main factor in the increase in Asia is an increase in the number of the applications in China in this period.

Furthermore, among 35 technology fields the largest share in the total of published patent applications worldwide in 2017 (3.199 million) was computer technology (7.3%), followed by electrical machinery/apparatus/energy (6.7%), measurement technology (5.1%), and medical technology (4.6%): the first two of these are in the field of electrical engineering and the other two are in the field of instruments.

The effects of R&D on firms themselves and on the economy in general can be measured by the returns on R&D. There have been many empirical studies to estimate the private and social returns on R&D. Sveikauskas (2007) reviews the estimates of the private and social returns on R&D shown in the previous studies. The private return on R&D has generally been estimated by comparing productivity growth or profitability in different firms with R&D expenditures or the growth of the research stock within these firms (Sveikauskas 2007). On the other hand, in order to estimate the returns to an industry or a national economy or even the returns to the world economy, the spillover effects of R&D and complementary investments have to be taken into account (Sveikauskas 2007). An example of complementary investment is that when a new computer is introduced, purchasing firms must deploy considerable resources to use the new equipment effectively. According to Sveikauskas (2007), the estimates of the private return on R&D in the previous studies range from 10% to 43%, whereas those of the social return range from 11% to 147%. He concludes that the private returns of 25% and the social returns of 65% seem reasonable. Since the social rates of returns include returns due to spillovers of knowledge in addition to private returns, the estimates become two or three times as large as the private rates of returns.¹³

Spillover channels of R&D performance are various. They are, for example, disclosure of patents, reverse engineering of newly developed products, movements of researchers and technical experts among organizations, research exchanges among them, industrial espionage, outsourcing, FDI, and so on.¹⁴

Two major channels of international diffusion of knowledge are international trade and FDI (Keller 2004, 2010). A number of studies confirm significant spillovers of knowledge through trade, but the empirical findings on spillover effects through FDI vary substantially. Spillovers of knowledge from foreign investors to local firms in the same sector are called horizontal spillovers, while knowledge spillovers from foreign investors to local firms in upstream and downstream sectors are called vertical spillovers. A large number of existing studies show that vertical spillovers from FDI tend to be positive and large, whereas horizontal ones from FDI are almost negligible (Keller 2004). However, the results vary substantially across countries, sectors, and estimation methods.¹⁵ Therefore, spillovers from FDI have attracted great attention from economists, and a number of studies using meta-analysis approaches have been conducted to figure out what factors cause differences in estimates (Görg and Strobl 2001; Havránek and Iršová 2011; Iršová and Havránek 2013; Meyer and Sinani 2009). With regard to horizontal spillovers, Iršová and Havránek (2013) find that investments through joint ventures between foreign investors and domestic firms tend to bring more positive spillovers than full foreign-ownership ones. The degree of technology gap is also important. They find that spillovers get smaller when the technology gap between foreign investors and domestic firms is too large. Moreover, Meyer and Sinani (2009) show that horizontal spillovers are related to a U-shaped form to the host economy’s level of development in terms of income, institutional framework, and human capital. As for vertical spillovers, Havránek and Iršová (2011) find that spillovers tend to be larger for host economies open to international trade and underdeveloped financial systems. In addition, greater spillovers seem to be generated by FDI from more distant countries with slight technological advantages over domestic firms.

The majority of the existing empirical studies on international diffusion of knowledge have analyzed the spillover effects of foreign knowledge on the productivity of domestic firms, but there is another type of equally important spillover effect. That is, the R&D activities of some firms or researchers may benefit from spillovers of knowledge that originated from innovation or the outcomes of R&D by other firms or researchers. To distinguish these two types of spillover effects, we call the former type “international productivity spillovers” and the latter type “international technology spillovers.”¹⁶ Although both types of spillovers capture flows of knowledge across countries, the exact effects differ. In this book we focus on the latter type. In the industrial organization literature, there are a large number of theoretical studies on the latter type of spillover among firms located in the same country (e.g., d’Aspremont and Jacquemin 1988; Haruna and Goel 2017; Kamien et al. 1992; Leahy and Neary 1997; Suzumura 1992). Moreover, theoretical studies on international technology spillovers and policies to address these spillovers include Goel and Haruna (2011), Haruna and Goel (2015), Leahy and Neary (1999), Neary and Leahy (2000), Neary and O’Sullivan (1999), and Qiu and Tao (1998). As we will explain in Chaps. 5–7, there are a number of existing empirical studies that investigate international technology spillovers, such as Branstetter (2006), Cappelli and Montobbio (2020), Haruna et al. (2010), Hu and Jaffe (2003), Jinji et al. (2013, 2015, 2019a), Li (2014), MacGarvie (2006), Mancusi (2008), Peri (2005), and Singh (2007).

1.2 Overview of the Book

In this book we explore interactions among deep integration, global firms, and technology spillovers. The structure of the book is as follows. In Chap. 2 we illustrate the trend of regional trade integration by distinguishing deep integration from shallow one. We clarify what deep regional integration means and discuss how we can measure shallow and deep integration.

In Chaps. 3 and 4 we focus on the behavior of global firms using micro data on Japanese MNEs. Specifically, we empirically investigate how firm performance, such as productivity and Tobin’s q, affects the choice of globalization mode. As for the modes of globalization, we consider export, FDI, and foreign outsourcing (FO). In Chaps. 5 and 6, we examine the relationship between global firms and technology spillovers. Specifically, we consider how trade patterns influence technology spillovers among countries/regions in Chap. 5. First of all, trade patterns can be classified into one-way trade (OWT or inter-industry trade) and two-way trade (or intra-industry trade: IIT). IIT is, furthermore, decomposed into horizontal intra-industry trade (HIIT) and vertical intra-industry trade (VIIT). These trade patterns arise from the behavior of heterogeneous firms. On the other hand, technology spillovers among countries/regions are measured by citations of patents. As we will explain later, patent citation data are used as a direct measure of technology spillovers (Hall et al. 2001).

Chapter 6 is devoted to the analysis of how FDI promotes technology spillovers. Employing detailed data on Japanese MNEs and their foreign affiliates, we measure the types of FDI (i.e., vertical and horizontal FDI). Then, we examine how differences in the types of FDI affect the degree of technology spillovers between Japanese MNEs and their host economies. Again, we utilize patent citation data.

Finally, we analyze the relationship between deep integration and technology spillovers in Chap. 7. Deep RTAs include a number of provisions that may directly affect flows of knowledge among countries/regions. We examine which aspects of deep integration contribute to enhance technology spillovers among members of RTAs.

To consider how deep integration facilitates technology spillovers, we need to take various channels of technology spillovers into consideration. Deep integration stimulates globalized activities of firms in a number of ways, which facilitates technology spillovers through various channels. For example, as for international trade, firms can obtain the necessary technology information from imported goods by dismantling them and making imitations. Moreover, when enterprises establish their operations overseas, this causes the transfer of production and business management technology to local enterprises in the host country. Experts, employees, and managers that are locally hired by the enterprises make their technologies, know-how, and knowledge with respect to their production and process management diffuse through their movements. In contrast, offshoring creates more direct diffusion of technological information. A representative example of offshoring is electronics manufacturing services (EMS) in Taiwan. In recent years, the number of firms without fabrication facilities (“fabless companies”) has increased. They specialize their activities in the design, development, and sales of products, and outsource their manufacturing. Apple Inc., which commissions the manufacturing of iPhone to enterprises in Taiwan and China, is one of the well-known fabless enterprises. Local enterprises overseas can acquire and accumulate information on the content and ways of manufacturing products through such commissioning.

Chapter 8 concludes and provides policy implications. Furthermore, it discusses some issues for future research in connection with this book.

Let us now look at each of these chapters in more detail.

Chapter 2 “The Trend of Deep Regional Integration”

The main purpose of this chapter is to illustrate the current trend of regional trade integration by distinguishing deep integration from shallow one and to shed light on the causes and impacts of recent deep regional integration. We explain what shallow and deep RTAs mean and why countries have recently pursued deep integration. Next, we argue how we can measure the degree of deep RTAs and what data are available to analyze the content and effects of deep RTAs. Moreover, we examine the state of deep RTAs in the world generally and in the Asia-Pacific region specifically.

The concept of shallow and deep integration is originally proposed by Lawrence (1996). Shallow integration is simply trade liberalization involving the removal of trade barriers. By contrast, deep integration “moves beyond the removal of border barriers” (Lawrence 1996, p. 8). Deep RTAs contain a variety of provisions including those on investment, labor, the environment, and intellectual property rights (IPR).

Horn et al. (2010) identify 52 policy areas covered by RTAs and classify them into two groups, i.e., WTO-plus (WTO+) and WTO-extra (WTO–X). The WTO+ group includes 14 provisions and the WTO-X group includes 38 ones. Limão (2016) recategorizes the WTO+ and WTO–X policy areas from the viewpoints of the depth and breadth of RTAs. The depth of RTAs measures the degree of bilateral economic cooperation. As for the definition of the depth of an RTA, he proposes that the depth of RTAs is measured by four categories of policy areas in the WTO+ and WTO–X groups: (a) import tariffs, (b) non-tariff barriers (NTBs), (c) behind-the-border policies (BBPs), and (d) other policies (OPs). On the other hand, the breadth of RTAs measures how wide the coverage of policy areas is. Policy areas are classified by (i) the type of trade (goods/services), (ii) technology (innovation/spillovers/IPR), and (iii) factors of production (capital/labor). As the definition of the breadth of an RTA, he proposes that the breadth of RTAs is measured by five categories: (a) services, (b) technology, (c) investment/capital, (d) labor, and (e) non-economic policies (NEPs).

We examine the characteristics of deep RTAs in the Asia-Pacific region. The contents of RTAs signed by the Association of South-East Asian Nations (ASEAN) countries, China and Japan are fairly different in their depth and breadth. Specifically, there are significant differences in the NTB and BBP areas in the depth measure and in the investment/capital and labor provisions in the breadth one. The degree of deep RTAs in the Asia-Pacific region is still not so high in both the OPs in the depth measure and the NEPs in the breadth one, compared with those in OP areas.

Chapter 3 “Which Aspect of Firm Performance is Important for the Choice of Globalization Mode?”

In this chapter we attempt to compare the effects of various measures of firm performance on firms’ globalization activities. Japanese firm-level data (covering the period 1994–1999) are used. Information on corporate balance sheets and patent applications are included in the data. Then, following Jinji et al. (2019b), we estimate the degree of engagement in each globalization mode by calculating the ratio of a mode of globalization activities such as export, FDI, and FO to the domestic sales of headquarters companies. Besides, we estimate the relative choice of the globalization mode by taking the ratio of the volume of direct export by the headquarters company to FDI (i.e., sales of foreign affiliates) and the ratio of costs of FO to FDI. As the measures of firm performance, we use three variables. First, labor productivity (LP) is used as a measure of productivity. Further, we employ two measures to capture the importance of the knowledge-capital intensity: one measure, as the second variable, is Tobin’s q (Tobin 1969) estimated by a simple approximation version; and the other, as the third variable, is the intangible asset intensity, which is the ratio of intangible to tangible assets. Intangible assets include patents, copyrights, trademarks, trade names, goodwill, and other items that lack physical substance and provide long-term benefits to the company. By using the stock of patent applications as a direct measure of intangible assets we regress the indexes of a firm’s choice of globalization mode on these variables.

By using Japanese firm-level data, we investigate empirically which measure of performance is important when a firm chooses one from various modes of globalization activity. As a result, it is found that an increase in LP or Tobin’s q motivates a firm to engage in export and FDI more, but does not enhance the engagement in FO. Secondly, it is found by using quantile regression that a difference in LP is important to a choice between exporting and FDI, but not to a choice between FDI and FO. In contrast, a difference in Tobin’s q is important to a choice between FDI and FO, but not to the one between exporting and FDI. Interestingly, a difference in the intangible asset intensity is important to a choice between exporting and FDI as well as to the one between FDI and FO.

Chapter 4 “Does Tobin’s q Matter for a Firm Choice of Globalization Mode?”

Previous theoretical research on the relationship between the productivities of firms and their globalization modes includes, for example, the following: Melitz (2003) presents a model in which the most productive firms export goods to foreign markets, whereas less productive firms supply goods only to their domestic market; and Helpman et al. (2004) extend the framework of Melitz and predict that only the most productive firms find it profitable to serve foreign markets via FDI and that medium productivity firms serve foreign markets through exports (the HMY prediction).

We attempt to sort firms into three modes of globalization by Tobin’s q (Tobin 1969). Our study is motivated by a theoretical analysis by Chen et al. (2012), who examine how the relative importance of knowledge capital over physical capital affects a firm’s choice between FDI and FO for offshore production. They then show that firms with a higher physical-capital intensity tend to choose FO, whereas firms with a higher knowledge-capital intensity tend to conduct FDI. An interesting testable hypothesis is obtained from this result: firms with a high Tobin’s q tend to conduct FDI, whereas firms with a low Tobin’s q tend to choose FO (the CHM hypothesis). Given that the book value of capital reflects only physical assets, a firm with a higher knowledge-capital intensity will have a higher Tobin’s q, because the firm’s market value reflects both knowledge-based and physical assets.

We employ detailed Japanese firm-level data (covering the period 1994–1999) to sort firms into three modes of globalization by Tobin’s q. The data include information on sales, employment, capital, R&D expenditure, direct exports, the costs of domestic production and FO of the companies headquartered in Japan, and the sales of their foreign affiliates. Corporate balance sheet data are also included. The advantage of the data over previous studies allows us to recognize not only whether a firm engages in a particular globalization activity among exports, FDI, and FO, but also the extent to which it is involved in that activity. By utilizing the feature of the data, we construct indexes to measure the relative choice of globalization mode through calculating both ratios of the costs of FO to the total FDI and of the volume of direct exports by the headquarters company to horizontal FDI. We then regress these indexes of globalization activity on Tobin’s q, of which measurement is based on the simple approximation proposed. To demonstrate how sorted patterns by Tobin’s q are different from those by a firm’s productivity, we regress the indexes of globalization activity on the total factor productivity (TFP) of each individual firm. Our analysis focuses mainly on firms engaging in multiple globalization modes and attempts to reveal whether Tobin’s q (and TFP) motivates the firms to select more FDI relative to FO or exports.

The main findings are as follows. Both quantile and endogenous quantile regressions indicate that an increase in Tobin’s q significantly reduces the ratio of FO to the total FDI across different quantiles, which strongly supports the CHM hypothesis that an MNE reduces its ratio of FO to FDI, as Tobin’s q increases. Besides, Tobin’s q has a positive effect on the ratio of exports to horizontal FDI at some quantiles, but is not strong. This implies that the imperfect contractibility of knowledge capital and a higher cost of technology transfer actually matter for knowledge-capital intensive firms. These effects of Tobin’s q on a firm’s choice of globalization mode are apparently different from those of TFP. An increase in TFP motivates a firm to enhance its engagement in horizontal FDI relative to exports, which supports the HMY prediction and concurs with existing empirical findings, but a difference in TFP does not significantly affect the choice of a firm between FDI and outsourcing.

Chapter 5 “Trade Patterns and International Technology Spillovers: Theory and Evidence from Japanese and European Patent Citations”

The international trade of goods and services is considered to be a major channel of technology spillovers. A simple explanation for this is that firms in an importing country can obtain information on advanced technology by, for example, reverse engineering of imported goods, patent information, and the movement of business persons. However, the relationship between bilateral trade patterns (such as inter-industry and intra-industry trade) and international technology spillovers has received little attention in the literature. It is intuitively conceivable that the flow of international knowledge could be different, depending on whether a good is only imported, exported, or both imported and exported. We have a look at the relationship between trade patterns and technology spillovers. It is worth considering the relationship from a theoretical point of view. To address this issue, we develop a two-country model of monopolistic competition with quality differentiation, in which inter- and intra-industry trade patterns endogenously arise. Our model is an extension of Melitz and Ottaviano (2008). It is assumed that firms randomly draw their product quality and hence are heterogeneous in product quality even if their productivity is identical. We investigate how technology spillovers are associated with trade patterns.

One theoretical feature of our model is that it can explain OWT, HIIT, and VIIT in a unified framework. Then our framework can provide the three testable hypotheses: the first hypothesis is that technology spillovers are larger when the trade pattern between the two countries is HIIT than when it is VIIT; the second one is whether or not technology spillovers from the country exporting higher quality products to that exporting lower quality ones are larger than those in the opposite direction is ambiguous when the trade pattern is VIIT; and the third one is that technology spillovers are lower when the trade pattern is inter-industry trade (i.e., OWT) than when it is VIIT.

Secondly, we empirically examine whether the three hypotheses theoretically derived hold. Our empirical analysis in this chapter complements Jinji et al. (2015): we use patent citation data at Japanese and European patent offices, whereas Jinji et al. (2015) employ US patent data. Our estimation results basically confirm the predictions of our theoretical model. An increase in the shares of HIIT and VIIT has a significantly positive effect on international technology spillovers. In addition, HIIT has a larger effect on them than VIIT does. In contrast, the relative magnitudes of technology spillovers between the country exporting high quality products and the country exporting low quality ones under VIIT are generally ambiguous. It is derived that the effect of OWT on technology spillovers tends to be much weaker than that of other trade patterns. Finally, we concluded that intra-industry trade plays a significant role in technology spillovers.

Chapter 6 “Vertical versus Horizontal Foreign Direct Investment and Technology Spillovers”

In this chapter we attempt to identify how MNEs’ activities in terms of horizontal and vertical FDI affect technology spillovers between themselves and host countries. Then, we combine the Japanese firm-level data on the business activities of Japanese MNEs’ foreign affiliates and the data on the patent citations at the US patent office between MNEs and their host countries. We now define a measure of “pure horizontal FDI” as the extent to which affiliates’ purchases of intermediate inputs and sales of final goods are concentrated in the local market and a measure of “pure vertical FDI” as the extent to which their purchases of intermediate inputs and sales of final goods are linked to the home country. We then estimate how the two types of FDI affect both technology spillovers from Japanese MNEs to the host country and from the host country to them by utilizing a negative binomial model. Moreover, to deal with a potential endogeneity problem we employ an endogenous switching model.

We obtain interesting results concerning technology spillovers under vertical and horizontal FDI through empirical analysis. First, an increase in the degree of pure vertical FDI has significantly positive effects on technology spillovers captured by patent citations when technologically advanced economies host Japanese MNEs (call this “result V”). Technology spillovers occur in both directions between the MNEs and their host countries. These positive effects of pure vertical FDI on them are robust for different specifications, and partially vertical FDI has significantly positive effects on technology spillovers from the (high-income) host countries to the MNEs. By contrast, an increase in the degree of pure horizontal FDI has no significant effects or significantly negative effects on technology spillovers between the MNEs and their host countries (call this “result H”). Partially horizontal FDI has significantly positive effects on them from the MNEs to the (high-income) host countries, but this result is not robust for different estimations. It is concluded from these results that pure vertical FDI plays a dominant role in the technology spillovers in both directions between the MNEs and the high-income host countries.

To explain the observed results between the structure of FDI and technology spillovers, we develop a simple partial-equilibrium model of FDI and technology spillovers among developed countries, in which differentiated goods are produced in three stages. The market is characterized by monopolistic competition. Depending on parameter values, firms may have an incentive to engage in horizontal or vertical FDI. Given the same factor costs in the two countries, there is no possibility of vertical FDI in the usual sense. However, vertical FDI does occur if there are technology gaps in some production stages between the countries and if firms can take advantage of the superior technology of the foreign country by fragmenting its production process and conducting some intermediate production in the foreign country. This explains result V observed in the empirical analysis. The technology gaps are the source of technology spillovers through FDI. Technology spillovers may occur in one way or two ways if firms engage in vertical FDI, depending on how the three production stages are located in the two countries. It is also shown that horizontal FDI does not necessarily induce technology spillovers, because it is mainly motivated by saving transportation costs and appears even in the absence of technological difference. This result corresponds with result H.

Chapter 7 “Do Deep Regional Trade Agreements Enhance International Technology Spillovers ?: Depth, Breadth, and Heterogeneity”

A rapid proliferation of RTAs has been observed during the last two decades. RTAs are primarily aimed at expanding trade in goods by reducing tariffs on imports and removing non-tariff barriers, but many recent RTAs pursue deeper integration, and include liberalization of investment and harmonization of IPR protection policy. It seems that RTAs affect the diffusion of knowledge (i.e., technology spillovers) across countries. We empirically investigate this issue.

We use patent citation data as a proxy for technology spillovers. Previous research shows that technology spillovers measured by patent citations decrease as geographical distance extends, but has paid little attention to the effects of “economic distance” on the localization of technology spillovers. Economic distance is a measure of proximity between two locations in an economic sense, which is affected by not only geographical distance but also other factors such as membership of RTAs, infrastructure, a transportation mode, and public policy. A given geographical distance between two countries (or regions) is constant, although the economic distance can vary, depending on such factors. Therefore, economic distance seems to be more meaningful to a measurement of technology spillovers than geographical distance. In particular, the membership of the same RTA or any other organization to facilitate international trade of goods will affect the economic distance between two countries and be of importance for the localization of technology spillovers.

Peri (2005) and Jinji et al. (2019a) have investigated the effects of RTAs on technology spillovers. Using a sample of 18 countries with 147 subnational regions in Western Europe and North America for the period of 1975–1996, Peri (2005) estimates a gravity-like model to examine the effects of several resistance factors on patent citations. He shows that regional, national, and linguistic borders have a significantly negative effect on technology spillovers, whereas the effect of “trade blocs” on them is insignificant. By contrast, Jinji et al. (2019a) find a significantly positive effect of RTAs on technology spillovers for the sample of 114 countries/regions during 1991–2007.

In comparison to these studies, we conduct a more comprehensive analysis of the effects of RTAs on technology spillovers by extending the sample to 243 countries/regions and the coverage of RTAs to 110. Specifically, we extend the research of Jinji et al. (2019a). For example, we construct a panel for 11,667 pairs of the citing and cited countries/regions from the sample of 243 countries/regions for 25 years from 1991 to 2015 by patent application and citation data from the US patent office.

We focus both on the impacts of the depth and breadth of RTAs and the heterogeneous effects of individual RTAs on international technology spillovers. With regard to the depth of RTAs, we use indexes in the areas of tariffs, NTBs, BBPs, and OPs. As for the breadth of RTAs, we construct indexes for services, technology, investment/capital, labor, and NEP areas. On the other hand, the heterogeneous effects of individual RTAs are captured by various RTA dummies. In addition to the usual RTA dummy, we use separate dummy variables for RTAs signed by the United States and European countries, such as the North American Free Trade Agreement (NAFTA), the European Community (EC)/European Union (EU), RTAs with the United States, and FTAs with the EC/EU.

The main findings are as follows. First, we confirm the result of Jinji et al. (2019a) that RTAs increase international technology spillovers measured by bilateral patent citations. Second, we demonstrate that deep RTAs with higher coverage of policy areas in the depth categories and those with higher coverage of policy areas in the breadth categories both have positive effects on international technology spillovers. Third, we show that the NAFTA has a strongly positive effect on such spillovers in comparison with the effects of the EU and EU enlargement.